Search

This study aimed to develop and test an unbiased and rapid methodology to estimate the length of external arbuscular mycorrhizal fungal (AMF) hyphae in soil. The traditional visual gridline intersection (VGI) method, which consists in a direct visual examination of the intersections of hyphae with gridlines on a microscope eyepiece after aqueous extraction, membrane-filtration, and staining (e.g., with trypan blue), was refined. For this, (i) images of the stained hyphae were taken by using a digital photomicrography technique to avoid the use of the microscope and the method was referred to as "digital gridline intersection" (DGI) method; and (ii), the images taken in (i) were processed and the hyphal length was measured by using ImageJ software, referred to as the "photomicrography-ImageJ processing" (PIP) method. The DGI and PIP methods were tested using known grade lengths of possum fur. Then they were applied to measure the hyphal lengths in soils with contrasting phosphorus (P) fertility status. Linear regressions were obtained between the known lengths (Lknown) of possum fur and the values determined by using either the DGI (LDGI) (LDGI = 0.37 + 0.97 × Lknown, r2 = 0.86) or PIP (LPIP) methods (LPIP = 0.33 + 1.01 × Lknown, r2 = 0.98). There were no significant (P > 0.05) differences between the LDGI and LPIP values. While both methods provided accurate estimation (slope of regression being 1.0), the PIP method was more precise, as reflected by a higher value of r2 and lower coefficients of variation. The average hyphal lengths (6.5-19.4 m g-1) obtained by the use of these methods were in the range of those typically reported in the literature (3-30 m g-1). Roots growing in P-deficient soil developed 2.5 times as many hyphae as roots growing in P-rich soil (17.4 vs 7.2 m g-1). These tests confirmed that the use of digital photomicrography in conjunction with either the grid-line intersection principle or image processing is a suitable method for the measurement of AMF hyphal lengths in soils for comparative investigations.

Background There are many types of hand tumors, and it is often difficult for imaging diagnosticians to make a correct diagnosis, which can easily lead to misdiagnosis and delay in treatment. Thus in this paper, we propose a deep neural network for diagnose on MR Images of tumors of the hand in order to better define preoperative diagnosis and standardize surgical treatment. Methods We collected MRI figures of 221 patients with hand tumors from one medical center from 2016 to 2019, invited medical experts to annotate the images to form the annotation data set. Then the original image is preprocessed to get the image data set. The data set is randomly divided into ten parts, nine for training and one for test. Next, the data set is input into the neural network system for testing. Finally, average the results of ten experiments as an estimate of the accuracy of the algorithm. Results This research uses 221 images as dataset and the system shows an average confidence level of 71.6% in segmentation of hand tumors. The segmented tumor regions are validated through ground truth analysis and manual analysis by a radiologist. Conclusions With the recent advances in convolutional neural networks, vast improvements have been made for image segmentation, mainly based on the skip-connection-linked encoder decoder deep architectures. Therefore, in this paper, we propose an automatic segmentation method based on DeepLab v3+ and achieved a good diagnostic accuracy rate.

The lensless optical fluid microscopy is of great significance to the miniaturization, portability and low cost development of cell detection instruments. However, the resolution of the cell image collected directly is low, because the physical pixel size of the image sensor is the same order of magnitude as the cell size. To solve this problem, this paper proposes a super-resolution scanning algorithm using a dual-line array sensor and a microfluidic chip. For dual-line array sensor images, the multi-group velocity and acceleration of cells flowing through the line array sensor are calculated. Then the reconstruction model of the super-resolution image is constructed with variable acceleration. By changing the angle between the line array image sensor and the direction of cell flow, the super-resolution image scanning and reconstruction are achieved in both horizontal and vertical directions. In addition, it is necessary to study the row by row extraction algorithm for cell foreground image. In this paper, the dual-line array sensor is implemented by adjusting the acquisition window of the image sensor with a pixel size of 2.2μm. When the tilt angle is 21 degrees, the equivalent pixel size is 0.79μm, improved 2.8 times, and after de-diffraction its average size error was 3.249%. As the angle decreases, the image resolution is higher, but the amount of information is less. This super-resolution scanning algorithm can be integrated on the chip and used with a microfluidic chip to realize on-chip instrument.

Multiscale geometric analysis (MGA) is not only characterized by multi-resolution, time-frequency localization, multidirectionality and anisotropy, but also outdoes the limitations of wavelet transform in representing high-dimensional singular data such as edges and contours. Therefore, researchers have been exploring new MGA-based image compression standards rather than the JPEG2000 standard. However, due to the difference in terms of the data structure, redundancy and decorrelation between wavelet and MGA, as well as the complexity of the coding scheme, so far, no definitive researches have been reported on the MGA-based image coding schemes. In addressing this problem, this paper proposes an image data compression approach using the hidden Markov model (HMM)/pulse-coupled neural network (PCNN) model in the contourlet domain. First, a sparse decomposition of an image was performed using a contourlet transform to obtain the coefficients that show the multiscale and multidirectional characteristics. An HMM was then adopted to establish links between coefficients in neighboring subbands of different levels and directions. An Expectation-Maximization (EM) algorithm was also adopted in training the HMM in order to estimate the state probability matrix, which maintains the same structure of the contourlet decomposition coefficients. In addition, each state probability can be classified by the PCNN based on the state probability distribution. Experimental results show that the HMM/PCNN -contourlet model proposed in this paper leads to better compression performance and offer a more flexible encoding scheme.

Image encryption is an effective method for protecting private images during communication. In this paper, a novel image encryption method is proposed based on a 5D hyperchaotic system. Since a 5D hyperchaotic system can generate more complex dynamic behavior than a low-dimensional system, it is used in this paper to generate pseudorandom number sequences. The generated sequences are processed to obtain new sequences. The randomness of the new sequences is improved by recombination and rearrangement. The experimental results and theoretical analysis show that the method possesses a large key space and can resist differential attacks, statistical analysis, entropy analysis, clipping attacks and noise attacks. Therefore, it is very secure and can be used for secure communication.

In the field of fashion design, designing garment image according to texture is actually changing the shape of texture image, and image-to-image translation based on Generative Adversarial Network (GAN) can do this well. This can help fashion designers save a lot of time and energy. GAN-based image-to-image translation has made great progress in recent years. One of the image-to-image translation models--StarGAN, has realized the function of multi-domain image-to-image translation by using only a single generator and a single discriminator. This paper details the use of StarGAN to complete the task of garment design. Users only need to input an image and a label for the garment type to generate garment images with the texture of the input image. However, it was found that the quality of the generated images is not satisfactory. Therefore, this paper introduces some improvements on the structure of the StarGAN generator and the loss function of StarGAN, and a model was obtained that can be better applied to garment design. It is called GD-StarGAN. This paper will demonstrate that GD-StarGAN is much better than StarGAN when it comes to garment design, especially in texture, by using a set of seven categories of garment datasets.

Color-based image segmentation classifies pixels of digital images in numerous groups for further analysis in computer vision, pattern recognition, image understanding, and image processing applications. Various algorithms have been developed for image segmentation, but clustering algorithms play an important role in the segmentation of digital images. This paper presents a novel and adaptive initialization approach to determine the number of clusters and find the initial central points of clusters for the standard K-means algorithm to solve the segmentation problem of color images. The presented scheme uses a scanning procedure of the paired Red, Green, and Blue (RGB) color-channel histograms for determining the most salient modes in every histogram. Next, the histogram thresholding is applied and a search in every histogram mode is performed to accomplish RGB pairs. These RGB pairs are used as the initial cluster centers and cluster numbers that clustered each pixel into the appropriate region for generating the homogeneous regions. The proposed technique determines the best initialization parameters for the conventional K-means clustering technique. In this paper, the proposed approach was compared with various unsupervised image segmentation techniques on various image segmentation benchmarks. Furthermore, we made use of a ranking approach inspired by the Evaluation Based on Distance from Average Solution (EDAS) method to account for segmentation integrity. The experimental results show that the proposed technique outperforms the other existing clustering techniques by optimizing the segmentation quality and possibly reducing the classification error.

In imaging systems, image blurs are a major source of degradation. This paper proposes a parameter estimation technique for linear motion blur, defocus blur, and atmospheric turbulence blur, and a nonlinear deconvolution algorithm based on sparse representation. Most blur removal techniques use image priors to estimate the point spread function (PSF); however, many common forms of image priors are unable to exploit local image information fully. In this paper, the proposed method does not require models of image priors. Further, it is capable of estimating the PSF accurately from a single input image. First, a blur feature in the image gradient domain is introduced, which has a positive correlation with the degree of blur. Next, the parameters for each blur type are estimated by a learning-based method using a general regression neural network. Finally, image restoration is performed using a half-quadratic optimization algorithm. Evaluation tests confirmed that the proposed method outperforms other similar methods and is suitable for dealing with motion blur in real-life applications.

Glioma detection and segmentation is a challenging task for radiologists and clinicians. The research reported in this paper seeks to develop a better clinical decision support algorithm for clinicians diagnosis. This paper presents a probabilistic method for detection and segmentation between abnormal tissue regions and brain tumour (tumour core and edema) portions from Magnetic Resonance Imaging (MRI). A framework is constructed to learn structure of undirected graphical models that can represent the spatial relationships among variables and apply it to glioma segmentation. Compared with the pixel of image, the superpixel is more consistent with human visual cognition and contains less redundancy, thus, the superpixels are considered as the basic unit of structure learning and glioma segmentation scheme. l1-regularization techniques are applied to learn the appropriate structure for modeling graphical models. Conditional Random Fields (CRF) are used to model the spatial interactions among image superpixel regions and their measurements. A number of features including statistics features, the combined features from the local binary pattern as well as gray level run length, curve features, and fractal features were extracted from each superpixel. The features are then passed by l1-regularization to ensure a robust classification. The proposed method is compared with support vector machine and Fuzzy c-means to classify each superpixel into normal and abnormal tissue. The proposed system is tested for the presence of low grade as well as high grade glioma tumors on images collected from BRATS2013, BRATS2015 data set and Henan Provincial People's Hospital (HNPPH) data set. The experiments performed provides similarity between segmented and truth image up to 91.5% by correlation method.

In developing nations, many expanding cities are facing challenges that result from the overwhelming numbers of people and vehicles. Collecting real-time, reliable and precise traffic flow information is crucial for urban traffic management. The main purpose of this paper is to develop an adaptive model that can assess the real-time vehicle counts on urban roads using computer vision technologies. This paper proposes an automatic real-time background update algorithm for vehicle detection and an adaptive pattern for vehicle counting based on the virtual loop and detection line methods. In addition, a new robust detection method is introduced to monitor the real-time traffic congestion state of road section. A prototype system has been developed and installed on an urban road for testing. The results show that the system is robust, with a real-time counting accuracy exceeding 99% in most field scenarios.

Compression at a very low bit rate(≤0.5bpp) causes degradation in video frames with standard decoding algorithms like H.261, H.262, H.264, and MPEG-1 and MPEG-4, which itself produces lots of artifacts. This paper focuses on an efficient pre-and post-processing technique (PP-AFT) to address and rectify the problems of quantization error, ringing, blocking artifact, and flickering effect, which significantly degrade the visual quality of video frames. The PP-AFT method differentiates the blocked images or frames using activity function into different regions and developed adaptive filters as per the classified region. The designed process also introduces an adaptive flicker extraction and removal method and a 2-D filter to remove ringing effects in edge regions. The PP-AFT technique is implemented on various videos, and results are compared with different existing techniques using performance metrics like PSNR-B, MSSIM, and GBIM. Simulation results show significant improvement in the subjective quality of different video frames. The proposed method outperforms state-of-the-art de-blocking methods in terms of PSNR-B with average value lying between (0.7–1.9db) while (35.83–47.7%) reduced average GBIM keeping MSSIM values very close to the original sequence statistically 0.978.

To reduce metal artifacts, several sinogram inpainting-based metal artifact reduction (MAR) methods have been proposed where projection data within the metal trace region of the sinogram are treated as missing and subsequently estimated. However, these methods generally assume data truncation does not occur and all metal objects reside inside the field-of-view (FOV). For small FOV imaging, these assumptions are violated, and thus existing inpainting-based MAR methods would not be effective. In this paper, we propose a new MAR method to reduce metal artifacts effectively in the presence of data truncation. The main idea behind the proposed method is the synthesis of a sinogram, which is treated as the originally measured sinogram. First, an initial reconstruction step is performed to remove truncation artifacts. The next step consists of a forward projection of the small FOV image, replacing the original sinogram with the synthesized sinogram. The final step is the application of sinogram inpainting based MAR methods using the synthesized sinogram. Verification of the proposed method was performed for three situations: extended cardiac-torso (XCAT) simulation, clinical data, and experimental data. The proposed method was applied with linear MAR (LMAR) and normalized MAR (NMAR), and the performance of the proposed method was compared with that of the previous method. For quantitative evaluation, normalized mean squared error (NMSE) and structure similarity (SSIM) were used. The results show the effectiveness of the proposed method to reduce residual metal artifacts, which are present in the results obtained with the previous method. The evaluation results using NMSE and SSIM also indicate that the proposed method is more effective than the previous method.

The process of Nuclei detection in high-grade breast cancer images is quite challenging in the case of image processing techniques due to certain heterogeneous characteristics of cancer nuclei such as enlarged and irregularly shaped nuclei, highly coarse chromatin marginalized to the nuclei periphery and visible nucleoli. Recent reviews state that existing techniques show appreciable segmentation accuracy on breast histopathology images whose nuclei are dispersed and regular in texture and shape; however, typical cancer nuclei are often clustered and have irregular texture and shape properties. This paper proposes a novel segmentation algorithm for detecting individual nuclei from Hematoxylin and Eosin (H&E) stained breast histopathology images. This detection framework estimates a nuclei saliency map using tensor voting followed by boundary extraction of the nuclei on the saliency map using a Loopy Back Propagation (LBP) algorithm on a Markov Random Field (MRF). The method was tested on both whole-slide images and frames of breast cancer histopathology images. Experimental results demonstrate high segmentation performance with efficient precision, recall and dice-coefficient rates, upon testing high-grade breast cancer images containing several thousand nuclei. In addition to the optimal performance on the highly complex images presented in this paper, this method also gave appreciable results in comparison with two recently published methods-Wienert et al. (2012) and Veta et al. (2013), which were tested using their own datasets. Science Foundation Ireland SFI-ISCA (Science Foundation Ireland - International Strategic Cooperation Award) program

The restoration of the Poisson noisy images is an essential task in many imaging applications due to the uncertainty of the number of discrete particles incident on the image sensor. In this paper, we consider utilizing a hybrid regularizer for Poisson noisy image restoration. The proposed regularizer, which combines the overlapping group sparse (OGS) total variation with the high-order nonconvex total variation, can alleviate the staircase artifacts while preserving the original sharp edges. We use the framework of the alternating direction method of multipliers to design an efficient minimization algorithm for the proposed model. Since the objective function is the sum of the non-quadratic log-likelihood and nonconvex nondifferentiable regularizer, we propose to solve the intractable subproblems by the majorization-minimization (MM) method and the iteratively reweighted least squares (IRLS) algorithm, respectively. Numerical experiments show the efficiency of the proposed method for Poissonian image restoration including denoising and deblurring.

In dynamic Positron Emission Tomography (PET), an estimate of the radio activity concentration is obtained from a series of frames of sinogram data taken at ranging in duration from 10 seconds to minutes under some criteria. So far, all the well-known reconstruction algorithms require known data statistical properties. It limits the speed of data acquisition, besides, it is unable to afford the separated information about the structure and the variation of shape and rate of metabolism which play a major role in improving the visualization of contrast for some requirement of the diagnosing in application. This paper presents a novel low rank-based activity map reconstruction scheme from emission sinograms of dynamic PET, termed as SLCR representing Sparse/Low Rank Constrained Reconstruction for Dynamic PET Imaging. In this method, the stationary background is formulated as a low rank component while variations between successive frames are abstracted to the sparse. The resulting nuclear norm and l 1 norm related minimization problem can also be efficiently solved by many recently developed numerical methods. In this paper, the linearized alternating direction method is applied. The effectiveness of the proposed scheme is illustrated on three data sets.

Deep learning has achieved a great success in natural image classification. To overcome data-scarcity in computational pathology, recent studies exploit transfer learning to reuse knowledge gained from natural images in pathology image analysis, aiming to build effective pathology image diagnosis models. Since transferability of knowledge heavily depends on the similarity of the original and target tasks, significant differences in image content and statistics between pathology images and natural images raise the questions: how much knowledge is transferable? Is the transferred information equally contributed by pre-trained layers? To answer these questions, this paper proposes a framework to quantify knowledge gain by a particular layer, conducts an empirical investigation in pathology image centered transfer learning, and reports some interesting observations. Particularly, compared to the performance baseline obtained by random-weight model, though transferability of off-the-shelf representations from deep layers heavily depend on specific pathology image sets, the general representation generated by early layers does convey transferred knowledge in various image classification applications. The observation in this study encourages further investigation of specific metric and tools to quantify effectiveness and feasibility of transfer learning in future., Comment: Experimentation data correction

The use of image covariates to build a classification model has lots of impact in various fields, such as computer science, medicine, and so on. The aim of this paper is to develop an estimation method for logistic regression model with image covariates. We propose a novel regularized estimation approach, where the regularization is a combination of L1 regularization and Sobolev norm regularization. The L1 penalty can perform variable selection, while the Sobolev norm penalty can capture the shape edges information of image data. We develop an efficient algorithm for the optimization problem. We also establish a nonasymptotic error bound on parameter estimation. Simulated studies and a real data application demonstrate that our proposed method performs very well.

Super-resolution (SR) technology provides a far promising computational imaging approach in obtaining a high-resolution (HR) image (or image sequences) from observed multiple low-resolution (LR) images by incorporating complementary information. In this paper, a three-stage SR method is proposed to generate a HR image from infrared (IR) LR Images acquired with Unmanned Aerial Vehicle (UAV). The proposed method integrates a high-level image capturing process and a low-level SR process. In this integrated process, we incorporate UAV path optimization, sub-pixel image registration, and sparseness constraint into a computational imaging framework of a region of interest (ROI). To refine ROI complementary feathers, we design an optimal flight control scheme to acquire adequate image sequences from multi-angles. In particular, a phase correlation approach achieving reliable sub-pixel image feature matching is adapted, on the basis of which an effective sparseness regularization model is built to enhance the fine structures of the IR image. Unlike most traditional multiple-frame SR algorithms that mainly focus on signal processing and achieve good performances when using standard test datasets, the performed experiments with real-life IR sequences indicate the three-stage SR method can also deal with practical LR IR image sequences collected by UAVs. The experimental results demonstrate that the proposed method is capable of generating HR images with good performance in terms of edge preservation and detail enhancement.

In recent years, deep learning (DL) networks have been widely used in super-resolution (SR) and exhibit improved performance. In this paper, an image quality assessment (IQA)-guided single image super-resolution (SISR) method is proposed in DL architecture, in order to achieve a nice tradeoff between perceptual quality and distortion measure of the SR result. Unlike existing DL-based SR algorithms, an IQA net is introduced to extract perception features from SR results, calculate corresponding loss fused with original absolute pixel loss, and guide the adjustment of SR net parameters. To solve the problem of heterogeneous datasets used by IQA and SR networks, an interactive training model is established via cascaded network. We also propose a pairwise ranking hinge loss method to overcome the shortcomings of insufficient samples during training process. The performance comparison between our proposed method with recent SISR methods shows that the former achieves a better tradeoff between perceptual quality and distortion measure than the latter. Extensive benchmark experiments and analyses also prove that our method provides a promising and opening architecture for SISR, which is not confined to a specific network model.

This paper presents an improved method to estimate the blur parameters of motion deblurring algorithm for single image restoration based on the point spread function (PSF) in frequency spectrum. We then introduce a modification to the Radon transform in the blur angle estimation scheme with our proposed difference value vs angle curve. Subsequently, the auto-correlation matrix is employed to estimate the blur angle by measuring the distance between the conjugated-correlated troughs. Finally, we evaluate the accuracy, robustness and time efficiency of our proposed method with the existing algorithms on the public benchmarks and the natural real motion blurred images. The experimental results demonstrate that the proposed PSF estimation scheme not only could obtain a higher accuracy for the blur angle and blur length, but also demonstrate stronger robustness and higher time efficiency under different circumstances.

Currently, injection sites of probes, cannula, and optic fibers in stereotactic neurosurgery are typically located manually. This step involves location estimations based on human experiences and thus introduces errors. In order to reduce localization error and improve repeatability of experiments and treatments, we investigate an automated method to locate injection sites. This paper proposes a localization framework, which integrates a region-based convolutional network and a fully convolutional network, to locate specific anatomical points on skulls of rodents. Experiment results show that the proposed localization framework is capable of identifying and locatin bregma and lambda in rodent skull anatomy images with mean errors less than 300 μm. This method is robust to different lighting conditions and mouse orientations, and has the potential to simplify the procedure of locating injection sites.

The Yunnan-Guizhou Plateau (YGP) is a typical ecologically fragile region in southwest China. Water-erosion desertification (WED) is one of the most significant environmental and socio-economic issues on the YGP and has seriously restricted the socio-economic development of this region. However, the research on monitoring of the desertification trends in this region has been limited to long time-series Landsat imagery. The objectives of this research were to monitor the WED trends on the YGP using time-series Landsat imagery data from 1989 to 2016. In this paper, we present a multi-indicator rule-based method, which was used to map the WED on the YGP during this period. The results show that the addition of multiple indicators improved the WED classification accuracy to 90.61%. Overall, the following results were obtained by using the proposed method. (1) The slight desertification area on the YGP increased from 89,617.09 km2 in 1989 to 100,976.47 km2 in 2016 with an annual growth ratio (AGR) of 0.48%, the moderate desertification area increased from 80,276.65 km2 in 1989 to 90,768.39 km2 in 2016 with an AGR of 0.50%, and the severe desertification area increased from 8149.3 km2 in 1989 to 13,220.16 km2 in 2016 with an AGR of 2.39%. (2) The WED expansion on the YGP can be divided into three stages. Firstly, the total WED area increased slowly from 17.80×104 km2 in 1989 to 17.98×104 km2 in 2010 with an AGR of 0.05%. Then, the WED rapidly expanded from 17.98×104 km2 in 2010 to 20.28×104 km2 in 2013 with an AGR of 4.26%. Finally, the WED increased slightly from 20.28×104 km2 in 2013 to 20.50×104 km2 in 2016 with an AGR of 0.36%. The total areas of the different degrees of WED decreased in 1992, 1998, 2001, and 2004. (3) The driving factors of WED were analyzed based on the Geographically Weighted Regression (GWR) model. We found that precipitation, vegetation area, and gross domestic product have key roles in the processes of desertification reversion and development. However, the regression coefficients between WED and these factors exhibited considerable spatial variations. The regression coefficients of the key driving factors showed different spatial distributions based on the GWR model in the YGP. The research results can provide scientific reference information for the prevention and control of WED in the YGP.

PET is a popular medical imaging modality for various clinical applications, including diagnosis and image-guided radiation therapy. The low-dose PET (LDPET) at a minimized radiation dosage is highly desirable in clinic since PET imaging involves ionizing radiation, and raises concerns about the risk of radiation exposure. However, the reduced dose of radioactive tracers could impact the image quality and clinical diagnosis. In this paper, a supervised deep learning approach with a generative adversarial network (GAN) and the cycle-consistency loss, Wasserstein distance loss, and an additional supervised learning loss, named as S-CycleGAN, is proposed to establish a non-linear end-to-end mapping model, and used to recover LDPET brain images. The proposed model, and two recently-published deep learning methods (RED-CNN and 3D-cGAN) were applied to 10% and 30% dose of 10 testing datasets, and a series of simulation datasets embedded lesions with different activities, sizes, and shapes. Besides vision comparisons, six measures including the NRMSE, SSIM, PSNR, LPIPS, SUVmax and SUVmean were evaluated for 10 testing datasets and 45 simulated datasets. Our S-CycleGAN approach had comparable SSIM and PSNR, slightly higher noise but a better perception score and preserving image details, much better SUVmean and SUVmax, as compared to RED-CNN and 3D-cGAN. Quantitative and qualitative evaluations indicate the proposed approach is accurate, efficient and robust as compared to other state-of-the-art deep learning methods.

High-speed imaging equipment can be an expensive investment, especially when certain applications require custom solutions. In this paper, we present a low-cost high-speed prototype camera built on a low-end Zynq-7000 System-on-Chip (SoC) platform and off-the-shelf components with the aim of removing the entry barrier into various high-speed imaging applications. The camera is standalone (does not require a host computer) and can achieve 211 frames per second (fps) at its maximum resolution of 1280x1024, and up to 2329 fps at a 256x256 resolution. With a current cost of only several hundred dollars and resource utilization of ~5%, the open-source design's modularity and customizability allows users with sufficient hardware or programming experience to modify the camera to suit their needs, potentially driving the cost lower. This can be done by utilizing the large remaining programmable logic for custom image processing algorithms, creating user interface software on the CPU, attaching extensions through the peripheral Module connections, or creating custom carrier or daughter boards. The development and design of the camera is described and a figure-of-merit is presented to provide a value assessment of some available commercial high-speed cameras against which our camera is competitive. Finally, the camera was tested to record low frequency spatial vibration and was found to be useful in investigating phenotypes associated with aging in a leading animal model, the nematode (worm) Caenorhabditis elegans.

In this paper, a new class of rational quadratic/linear trigonometric Hermite functions with two shape parameters is proposed. Based on these Hermite functions, new improved first class of Side-Side (FCSS), second class of Side-Side (SCSS), first class of Side-Vertex (FCSV) and second class of Side-Vertex (SCSV) interpolation operators are proposed respectively, which can be used to construct C1 Coons surfaces over triangular domain. By altering the values of two shape parameters, the shape of the Coons surface patch can be adjusted flexibly, but without affecting the function values and partial derivatives of the boundaries. For constructing the triangular surface patches with the center of mass passing through a fixed point, we also give a center of mass function value control method, by which we can solve the corresponding shape parameter values. Moreover, we also apply these four improved interpolation operators to image interpolation. Compared with some widely used image interpolation methods, our methods achieve competitive performance.

A system for real-time ultrasound (US) elastography will advance interventions for the diagnosis and treatment of cancer by advancing methods such as thermal monitoring of tissue ablation. A multi-stream graphics processing unit (GPU) based accelerated normalized cross-correlation (NCC) elastography, with a maximum frame rate of 78 frames per second, is presented in this paper. A study of NCC window size is undertaken to determine the effect on frame rate and the quality of output elastography images. This paper also presents a novel system for Online Tracked Ultrasound Elastography (O-TRuE), which extends prior work on an offline method. By tracking the US probe with an electromagnetic (EM) tracker, the system selects in-plane radio frequency (RF) data frames for generating high quality elastograms. A novel method for evaluating the quality of an elastography output stream is presented, suggesting that O-TRuE generates more stable elastograms than generated by untracked, free-hand palpation. Since EM tracking cannot be used in all systems, an integration of real-time elastography and the da Vinci Surgical System is presented and evaluated for elastography stream quality based on our metric. The da Vinci surgical robot is outfitted with a laparoscopic US probe, and palpation motions are autonomously generated by customized software. It is found that a stable output stream can be achieved, which is affected by both the frequency and amplitude of palpation. The GPU framework is validated using data from in-vivo pig liver ablation; the generated elastography images identify the ablated region, outlined more clearly than in the corresponding B-mode US images.

This paper focuses on the application of deep learning (DL) in the diagnosis of coronavirus disease (COVID-19). The novelty of this work is in the introduction of optimized InceptionResNetV2 for COVID-19 (CO-IRv2) method. A part of the CO-IRv2 scheme is derived from the concepts of InceptionNet and ResNet with hyperparameter tuning, while the remaining part is a new architecture consisting of a global average pooling layer, batch normalization, dense layers, and dropout layers. The proposed CO-IRv2 is applied to a new dataset of 2481 computed tomography (CT) images formed by collecting two independent datasets. Data resizing and normalization are performed, and the evaluation is run up to 25 epochs. Various performance metrics, including precision, recall, accuracy, F1-score, area under the receiver operating characteristics (AUC) curve are used as performance metrics. The effectiveness of three optimizers known as Adam, Nadam and RMSProp are evaluated in classifying suspected COVID-19 patients and normal people. Results show that for CO-IRv2 and for CT images, the obtained accuracies of Adam, Nadam and RMSProp optimizers are 94.97%, 96.18% and 96.18%, respectively. Furthermore, it is shown here that for the case of CT images, CO-IRv2 with Nadam optimizer has better performance than existing DL algorithms in the diagnosis of COVID-19 patients. Finally, CO-IRv2 is applied to an X-ray dataset of 1662 images resulting in a classification accuracy of 99.40%.

A high-density crowd state is prone to cause large-scale crowd stampede accidents that seriously threaten the people’s security and property. The key to preventing crowd congestion is to accurately predict the location and time of crowd events, particularly when there is a high density of people. In this paper, the entropy theory is used to characterize the state of a crowded system. The theoretical entropy Sr and the actual entropy S of the crowd system are obtained according to the area occupied by the different crowd state The maximum entropy value and the actual entropy value under different conditions of the system are compared to judge the state of crowded extent. The results show that the model is practical and effective. According to the situation of the crowd, different management and evacuation measures are considered to prevent the occurrence of crowd accidents.

Image segmentation is a fundamental task in image processing and is still a challenging problem when processing images with high noise, low resolution and intensity inhomogeneity. In this paper, a weighted region-based level set method, which is based on the techniques of local statistical theory, level set theory and curve evolution, is proposed. Specifically, a new weighted pressure force function (WPF) is first presented to flexibly drive the closed contour to shrink or expand outside and inside of the object. Second, a faster and smoother regularization term is added to ensure the stability of the curve evolution and that there is no need for initialization in curve evolution. Third, the WPF is integrated into the region-based level set framework to accelerate the speed of the curve evolution and improve the accuracy of image segmentation. Experimental results on medical and natural images demonstrate that the proposed segmentation model is more efficient and robust to noise than other state-of-the-art models.

Using deformable models to register medical images can result in problems of initialization of deformable models and robustness and accuracy of matching of inter-subject anatomical variability. To tackle these problems, a novel model is proposed in this paper by compounding local invariant features and global deformable geometry. This model has four steps. First, a set of highly-repeatable and highly-robust local invariant features, called Key Features Model (KFM), are extracted by an effective matching strategy. Second, local features can be matched more accurately through the KFM for the purpose of initializing a global deformable model. Third, the positional relationship between the KFM and the global deformable model can be used to precisely pinpoint all landmarks after initialization. And fourth, the final pose of the global deformable model is determined by an iterative process with a lower time cost. Through the practical experiments, the paper finds three important conclusions. First, it proves that the KFM can detect the matching feature points well. Second, the precision of landmark locations adjusted by the modeled relationship between KFM and global deformable model is greatly improved. Third, regarding the fitting accuracy and efficiency, by observation from the practical experiments, it is found that the proposed method can improve 6~8% of the fitting accuracy and reduce around 50% of the computational time compared with state-of-the-art methods.

This paper studies the problem of the restoration of images corrupted by mixed Gaussian-impulse noise. In recent years, low-rank matrix reconstruction has become a research hotspot in many scientific and engineering domains such as machine learning, image processing, computer vision and bioinformatics, which mainly involves the problem of matrix completion and robust principal component analysis, namely recovering a low-rank matrix from an incomplete but accurate sampling subset of its entries and from an observed data matrix with an unknown fraction of its entries being arbitrarily corrupted, respectively. Inspired by these ideas, we consider the problem of recovering a low-rank matrix from an incomplete sampling subset of its entries with an unknown fraction of the samplings contaminated by arbitrary errors, which is defined as the problem of matrix completion from corrupted samplings and modeled as a convex optimization problem that minimizes a combination of the nuclear norm and the l(1)-norm in this paper. Meanwhile, we put forward a novel and effective algorithm called augmented Lagrange multipliers to exactly solve the problem. For mixed Gaussian-impulse noise removal, we regard it as the problem of matrix completion from corrupted samplings, and restore the noisy image following an impulse-detecting procedure. Compared with some existing methods for mixed noise removal, the recovery quality performance of our method is dominant if images possess low-rank features such as geometrically regular textures and similar structured contents; especially when the density of impulse noise is relatively high and the variance of Gaussian noise is small, our method can outperform the traditional methods significantly not only in the simultaneous removal of Gaussian noise and impulse noise, and the restoration ability for a low-rank image matrix, but also in the preservation of textures and details in the image.

Today, while many researchers focus on the improvement of the regularization term in IR algorithms, they pay less concern to the improvement of the fidelity term. In this paper, we hypothesize that improving the fidelity term will further improve IR image quality in low-dose scanning, which typically causes more noise. The purpose of this paper is to systematically test and examine the role of high-fidelity system models using raw data in the performance of iterative image reconstruction approach minimizing energy functional. We first isolated the fidelity term and analyzed the importance of using focal spot area modeling, flying focal spot location modeling, and active detector area modeling as opposed to just flying focal spot motion. We then compared images using different permutations of all three factors. Next, we tested the ability of the fidelity terms to retain signals upon application of the regularization term with all three factors. We then compared the differences between images generated by the proposed method and Filtered-Back-Projection. Lastly, we compared images of low-dose in vivo data using Filtered-Back-Projection, Iterative Reconstruction in Image Space, and the proposed method using raw data. The initial comparison of difference maps of images constructed showed that the focal spot area model and the active detector area model also have significant impacts on the quality of images produced. Upon application of the regularization term, images generated using all three factors were able to substantially decrease model mismatch error, artifacts, and noise. When the images generated by the proposed method were tested, conspicuity greatly increased, noise standard deviation decreased by 90% in homogeneous regions, and resolution also greatly improved. In conclusion, the improvement of the fidelity term to model clinical scanners is essential to generating higher quality images in low-dose imaging.

Microscopic examination of peripheral blood plays an important role in the field of diagnosis and control of major diseases. Peripheral leukocyte recognition by manual requires medical technicians to observe blood smears through light microscopy, using their experience and expertise to discriminate and analyze different cells, which is time-consuming, labor-intensive and subjective. The traditional systems based on feature engineering often need to ensure successful segmentation and then manually extract certain quantitative and qualitative features for recognition but still remaining a limitation of poor robustness. The classification pipeline based on convolutional neural network is of automatic feature extraction and free of segmentation but hard to deal with multiple object recognition. In this paper, we take leukocyte recognition as object detection task and apply two remarkable object detection approaches, Single Shot Multibox Detector and An Incremental Improvement Version of You Only Look Once. To improve recognition performance, some key factors involving these object detection approaches are explored and the detection models are generated using the train set of 14,700 annotated images. Finally, we evaluate these detection models on test sets consisting of 1,120 annotated images and 7,868 labeled single object images corresponding to 11 categories of peripheral leukocytes, respectively. A best mean average precision of 93.10% and mean accuracy of 90.09% are achieved while the inference time is 53 ms per image on a NVIDIA GTX1080Ti GPU.

The paper presents the results of a quantitative analysis of speeches by Nobel laureates in the sciences (Physics, Chemistry, Medicine) at the Prize gala dinner throughout the whole history of the Prize, 1901–2018. The results outline key themes and historical trends. A dominant theme, common to most speeches, is the exaltation of science as a profession by the laureate. Since the 1970s, especially in chemistry, this element becomes more domain-specific and less related to science in general. One could speculate whether this happens chiefly in chemistry because its area of activity has been perceived to be at risk of erosion from competing fields (e.g. physics, biology). Over time, speeches become more technical, less ceremonial and more lecture-oriented. Emphasis on broad, beneficial impact of science for humanity and mankind (as emphasised in Nobel’s will) is more present in laureates’ speeches during the first half of the XXth century, while its relevance clearly declines during the last decades. Politics and its relationship with science is also a relevant topic in Nobel speeches. Particularly between the two World Wars, science is seen as terrain where nationalistic stances and fights among nations could actually find a context for peaceful competition and even cooperation.

This paper attempts to design an automated, efficient and intelligent tomato grading method that facilitates the graded selling of the fruit. Based on machine vision, the color images of tomatoes with different morphologies were studied, and the color, shape and size were selected as the key features. On this basis, an automated grading classifier was created based on the surface features of tomatoes, and a grading platform was set up to verify the effect of the classifier. Specifically, the Hue value distributions of tomatoes with different maturities were investigated, and the Hue value ranges were determined for mature, semi-mature and immature tomatoes, producing the color classifier. Next, the first-order Fourier descriptor (1D- FD) was adopted to describe the radius sequence of tomato contour, and an equation was established to compute the irregularity of tomato contour, creating the shape classifier. After that, a linear regression equation was constructed to reflect the relationship between the transverse diameters of actual tomatoes and tomato images, and a classifier between large, medium and small tomatoes was produced based on the transverse diameter. Finally, a comprehensive tomato classifier was built based on the color, shape and size diameters. The experimental results show that the mean grading accuracy of the proposed method was 90.7%. This means our method can achieve automated real-time grading of tomatoes.

Although successful detection of malignant tumors from histopathological images largely depends on the long-term experience of radiologists, experts sometimes disagree with their decisions. Computer-aided diagnosis provides a second option for image diagnosis, which can improve the reliability of experts' decision-making. Automatic and precision classification for breast cancer histopathological image is of great importance in clinical application for identifying malignant tumors from histopathological images. Advanced convolution neural network technology has achieved great success in natural image classification, and it has been used widely in biomedical image processing. In this paper, we design a novel convolutional neural network, which includes a convolutional layer, small SE-ResNet module, and fully connected layer. We propose a small SE-ResNet module which is an improvement on the combination of residual module and Squeeze-and-Excitation block, and achieves the similar performance with fewer parameters. In addition, we propose a new learning rate scheduler which can get excellent performance without complicatedly fine-tuning the learning rate. We use our model for the automatic classification of breast cancer histology images (BreakHis dataset) into benign and malignant and eight subtypes. The results show that our model achieves the accuracy between 98.87% and 99.34% for the binary classification and achieve the accuracy between 90.66% and 93.81% for the multi-class classification.

A new Interpolation based Reversible Data Hiding (IRDH) scheme is reported in this paper. For different applications of an IRDH scheme to the digital image, video, multimedia, big-data and biological data, the embedding capacity requirement usually varies. Disregarding this important consideration, existing IRDH schemes do not offer a better embedding rate-distortion performance for varying size payloads. To attain this varying capacity requirement with our proposed adaptive embedding, we formulate a capacity control parameter and propose to utilize it to determine a minimum set of embeddable bits in a pixel. Additionally, we use a logical (or bit-wise) correlation between the embeddable pixel and estimated versions of an embedded pixel. Thereby, while a higher range between an upper and lower limit of the embedding capacity is maintained, a given capacity requirement within that limit is also attained with a better-embedded image quality. Computational modeling of all new processes of the scheme is presented, and performance of the scheme is evaluated with a set of popular test-images. Experimental results of our proposed scheme compared to the prominent IRDH schemes have recorded a significantly better-embedding rate-distortion performance.

The recent progress in photon-counting detector technology using high-Z semiconductor sensors provides new possibilities for spectral x-ray imaging. The benefits of the approach to extract spectral information directly from measurements in the projection domain are very advantageous for material science studies with x-rays as polychromatic artifacts like beam-hardening are handled properly. Since related methods require accurate knowledge of all energy-dependent system parameters, we utilize an adapted semi-empirical model, which relies on a simple calibration procedure. The method enables a projection-based decomposition of photon-counting raw-data into basis material projections. The objective of this paper is to investigate the method's performance applied to x-ray micro-CT with special focus on applications in material science and non-destructive testing. Projection-based dual-energy micro-CT is shown to be of good quantitative accuracy regarding material properties such as electron densities and effective atomic numbers. Furthermore, we show that the proposed approach strongly reduces beam-hardening artifacts and improves image contrast at constant measurement time.

Diffusion MRI derives its contrast from MR signal attenuation induced by the movement of water molecules in microstructural environments. Associated with the signal attenuation is the reduction of signal-to-noise ratio (SNR). Methods based on total variation (TV) have shown superior performance in image noise reduction. However, TV denoising can result in stair-casing effects due to the inherent piecewise-constant assumption. In this paper, we propose a tight wavelet frame based approach for edge-preserving denoising of diffusion-weighted (DW) images. Specifically, we employ the unitary extension principle (UEP) to generate frames that are discrete analogues to differential operators of various orders, which will help avoid stair-casing effects. Instead of denoising each DW image separately, we collaboratively denoise groups of DW images acquired with adjacent gradient directions. In addition, we introduce a very efficient method for solving an l0 denoising problem that involves only thresholding and solving a trivial inverse problem. We demonstrate the effectiveness of our method qualitatively and quantitatively using synthetic and real data.

Image compression and image encryption are two essential tasks in image processing. The former aims to reduce the cost for storage or transmission of images while the latter aims to change the positions or values of pixels to protect image content. Nowadays, an increasing number of researchers are focusing on the combination of these two tasks. In this paper, we propose a novel joint image compression and encryption approach that integrates a quantum chaotic system, sparse Bayesian learning (SBL) and a bit-level 3D Arnold cat map, so-called QSBLA, for such a purpose. Specifically, the QSBLA consists of 6 stages. First, a quantum chaotic system is employed to generate chaotic sequences for subsequent compression and encryption. Second, as one method of compressive sensing, SBL is used to compress images. Third, an operation of diffusion is performed on the compressed image. Fourth, the compressed and diffused image is transformed into several bit-level cubes. Fifth, 3D Arnold cat maps are used to permute each bit-level cube. Finally, all the bit-level cubes are integrated and transformed into a 2D pixel-level image, resulting in the compressed and encrypted image. Extensive experiments on 8 publicly-accessed images demonstrate that the proposed QSBLA is superior or comparable to some state-of-the-art approaches in terms of several measurement indices, indicating that the QSBLA is promising for joint image compression and encryption.

With the increase of the camera resolution, the number of pixels contained in froth image is increased, which brings many challenges to image segmentation. Froth size and distribution are the important index in froth flotation. The segmentation of froth images is always a problem in building flotation model. In segmenting froth images, Otsu method is usually used to get a binary image for classification of froth images, this method can get a satisfactory segmentation result. However, each gray level is required to calculate each of the between-class variance, it takes a longer time in froth images with a large number of pixels. To solve this problem, an improved method is proposed in this paper. Most froth images have the pixel distribution characteristic that the gray histogram curve is a sawtooth shape. The proposed method uses polynomial to fit the curve of gray histogram and takes the characteristic of gray histogram's valley into consideration in Otsu method. Two performance comparison methods are introduced and used. Experimental comparison between Otsu method and the proposed method shows that the proposed method has a satisfactory image segmentation with a low computing time.

A recurrent criticism concerning the use of online social media data in political science research is the lack of demographic information about social media users. By employing a face-recognition algorithm to the profile pictures of Facebook users, the paper derives two fundamental demographic characteristics (age and gender) of a sample of Facebook users who interacted with the most relevant British parties in the two weeks before the Brexit referendum of 23 June 2016. The article achieves the goals of (i) testing the precision of the algorithm, (ii) testing its validity, (iii) inferring new evidence on digital mobilisation, and (iv) tracing the path for future developments and application of the algorithm. The findings show that the algorithm is reliable and that it can be fruitfully used in political and social sciences both to confirm the validity of survey data and to obtain information from populations that are generally unavailable within traditional surveys.

The traditional image mosaic result based on SIFT feature points extraction, to some extent, has distortion errors: the larger the input image set, the greater the spliced panoramic distortion. To achieve the goal of creating a high-quality panorama, a new and improved algorithm based on the A-KAZE feature is proposed in this paper. This includes changing the way reference image are selected and putting forward a method for selecting a reference image based on the binary tree model, which takes the input image set as the leaf node set of a binary tree and uses the bottom-up approach to construct a complete binary tree. The root node image of the binary tree is the ultimate panorama obtained by stitching. Compared with the traditional way, the novel method improves the accuracy of feature points detection and enhances the stitching quality of the panorama. Additionally, the improved method proposes an automatic image straightening model to rectify the panorama, which further improves the panoramic distortion. The experimental results show that the proposed method cannot only enhance the efficiency of image stitching processing, but also reduce the panoramic distortion errors and obtain a better quality panoramic result.

A number of recent studies have shown that cell shape and cytoskeletal texture can be used as sensitive readouts of the physiological state of the cell. However, utilization of this information requires the development of quantitative measures that can describe relevant aspects of cell shape. In this paper we develop a toolbox, TISMorph, that calculates a set of quantitative measures to address this need. Some of the measures introduced here have been used previously, while others are new and have desirable properties for shape and texture quantification of cells. These measures, broadly classifiable into the categories of textural, irregularity and spreading measures, are tested by using them to discriminate between osteosarcoma cell lines treated with different cytoskeletal drugs. We find that even though specific classification tasks often rely on a few measures, these are not the same between all classification tasks, thus requiring the use of the entire suite of measures for classification and discrimination. We provide detailed descriptions of the measures, as well as the TISMorph package to implement them. Quantitative morphological measures that capture different aspects of cell morphology will help enhance large-scale image-based quantitative analysis, which is emerging as a new field of biological data.

Optical imaging through biological samples is compromised by tissue scattering and currently various approaches aim to overcome this limitation. In this paper we demonstrate that an all optical technique, based on non-linear upconversion of infrared ultrashort laser pulses and on multiple view acquisition, allows the reduction of scattering effects in tomographic imaging. This technique, namely Time-Gated Optical Projection Tomography (TGOPT), is used to reconstruct three dimensionally the internal structure of adult zebrafish without staining or clearing agents. This method extends the use of Optical Projection Tomography to optically diffusive samples yielding reconstructions with reduced artifacts, increased contrast and improved resolution with respect to those obtained with non-gated techniques. The paper shows that TGOPT is particularly suited for imaging the skeletal system and nervous structures of adult zebrafish.

Due to the high sensitivity of diffusion tensor imaging (DTI) to physiological motion, clinical DTI scans often suffer a significant amount of artifacts. Tensor-fitting-based, post-processing outlier rejection is often used to reduce the influence of motion artifacts. Although it is an effective approach, when there are multiple corrupted data, this method may no longer correctly identify and reject the corrupted data. In this paper, we introduce a new criterion called "corrected Inter-Slice Intensity Discontinuity" (cISID) to detect motion-induced artifacts. We compared the performance of algorithms using cISID and other existing methods with regard to artifact detection. The experimental results show that the integration of cISID into fitting-based methods significantly improves the retrospective detection performance at post-processing analysis. The performance of the cISID criterion, if used alone, was inferior to the fitting-based methods, but cISID could effectively identify severely corrupted images with a rapid calculation time. In the second part of this paper, an outlier rejection scheme was implemented on a scanner for real-time monitoring of image quality and reacquisition of the corrupted data. The real-time monitoring, based on cISID and followed by post-processing, fitting-based outlier rejection, could provide a robust environment for routine DTI studies.

Diagnosis of retinal vascular diseases depends on ophthalmoscopic findings that most often occur after severe visual loss (as in vein occlusions) or chronic changes that are irreversible (as in diabetic retinopathy). Despite recent advances, diagnostic imaging currently reveals very little about the vascular function and local oxygen delivery. One potentially useful measure of vascular function is measurement of hemoglobin oxygen content. In this paper, we demonstrate a novel method of accurately, rapidly and easily measuring oxygen saturation within retinal vessels using in vivo imaging spectroscopy. This method uses a commercially available fundus camera coupled to two-dimensional diffracting optics that scatter the incident light onto a focal plane array in a calibrated pattern. Computed tomographic algorithms are used to reconstruct the diffracted spectral patterns into wavelength components of the original image. In this paper the spectral components of oxy- and deoxyhemoglobin are analyzed from the vessels within the image. Up to 76 spectral measurements can be made in only a few milliseconds and used to quantify the oxygen saturation within the retinal vessels over a 10-15 degree field. The method described here can acquire 10-fold more spectral data in much less time than conventional oximetry systems (while utilizing the commonly accepted fundus camera platform). Application of this method to animal models of retinal vascular disease and clinical subjects will provide useful and novel information about retinal vascular disease and physiology.

Accurate segmentation of brain magnetic resonance imaging (MRI) is an essential step in quantifying the changes in brain structure. Deep learning in recent years has been extensively used for brain image segmentation with highly promising performance. In particular, the U-net architecture has been widely used for segmentation in various biomedical related fields. In this paper, we propose a patch-wise U-net architecture for the automatic segmentation of brain structures in structural MRI. In the proposed brain segmentation method, the non-overlapping patch-wise U-net is used to overcome the drawbacks of conventional U-net with more retention of local information. In our proposed method, the slices from an MRI scan are divided into non-overlapping patches that are fed into the U-net model along with their corresponding patches of ground truth so as to train the network. The experimental results show that the proposed patch-wise U-net model achieves a Dice similarity coefficient (DSC) score of 0.93 in average and outperforms the conventional U-net and the SegNet-based methods by 3% and 10%, respectively, for on Open Access Series of Imaging Studies (OASIS) and Internet Brain Segmentation Repository (IBSR) dataset.

Preliminary diagnosis of fungal infections can rely on microscopic examination. However, in many cases, it does not allow unambiguous identification of the species by microbiologist due to their visual similarity. Therefore, it is usually necessary to use additional biochemical tests. That involves additional costs and extends the identification process up to 10 days. Such a delay in the implementation of targeted therapy may be grave in consequence as the mortality rate for immunosuppressed patients is high. In this paper, we apply a machine learning approach based on deep neural networks and Fisher Vector (advanced bag-of-words method) to classify microscopic images of various fungi species. Our approach has the potential to make the last stage of biochemical identification redundant, shortening the identification process by 2-3 days, and reducing the cost of the diagnosis.

ChronoMID builds on the success of cross-modal convolutional neural networks (X-CNNs), making the novel application of the technique to medical imaging data. Specifically, this paper presents and compares alternative approaches - timestamps and difference images - to incorporate temporal information for the classification of bone disease in mice, applied to micro-CT scans of mouse tibiae. Whilst much previous work on diseases and disease classification has been based on mathematical models incorporating domain expertise and the explicit encoding of assumptions, the approaches given here utilise the growing availability of computing resources to analyse large datasets and uncover subtle patterns in both space and time. After training on a balanced set of over 75000 images, all models incorporating temporal features outperformed a state-of-the-art CNN baseline on an unseen, balanced validation set comprising over 20000 images. The top-performing model achieved 99.54% accuracy, compared to 73.02% for the CNN baseline., Comment: 8 pages, 7 figures, 2 tables