Search

A variety of supervise learning methods are available for low-dose CT denoising in the sinogram domain. Traditional model observers are widely employed to evaluate these methods. However, the sinogram domain evaluation remains an open challenge for deep learning-based low-dose CT denoising. Since each lesion in medical CT images corresponds to a narrow sinusoidal strip in sinogram domain, here we proposed a transformer-based model observer to evaluate sinogram domain supervised learning methods. The numerical results indicate that our transformer-based model well-approximates the Laguerre-Gauss channelized Hotelling observer (LG-CHO) for a signal-known-exactly (SKE) and background-known-statistically (BKS) task. The proposed model observer is employed to assess two classic CNN-based sinogram domain denoising methods. The results demonstrate a utility and potential of this transformer-based observer model in developing deep low-dose CT denoising methods in the sinogram domain., 14 pages, 9 figures

Accurately identifying patient's status through medical images plays an important role in diagnosis and treatment. Artificial intelligence (AI), especially the deep learning, has achieved great success in many fields. However, more reliable AI model is needed in image guided diagnosis and therapy. To achieve this goal, developing a balanced, safe and robust model with a unified framework is desirable. In this study, a new unified model termed as automated multi-objective Mixer (AutoMO-Mixer) model was developed, which utilized a recent developed multiple layer perceptron Mixer (MLP-Mixer) as base. To build a balanced model, sensitivity and specificity were considered as the objective functions simultaneously in training stage. Meanwhile, a new evidential reasoning based on entropy was developed to achieve a safe and robust model in testing stage. The experiment on an optical coherence tomography dataset demonstrated that AutoMO-Mixer can obtain safer, more balanced, and robust results compared with MLP-Mixer and other available models.

Image quality assessment that aims at estimating the subject quality of images, builds models to evaluate the perceptual quality of the image in different applications. Based on the fact that the human visual system (HVS) is highly sensitive to structural information, the edge information extraction is widely applied in different IQA metrics. According to previous studies, the image gradient magnitude (GM) and the Laplacian of Gaussian (LOG) operator are two efficient structural features in IQA tasks. However, most of the IQA metrics achieve good performance only when the distorted image is totally registered with the reference image, but fail to perform on images with small translations. In this paper, we propose an FR-IQA model with the quadratic sum of the GM and the LOG signals, which obtains good performance in image quality estimation considering shift-insensitive property for not well-registered reference and distortion image pairs. Experimental results show that the proposed model works robustly on three large scale subjective IQA databases which contain a variety of distortion types and levels, and stays in the state-of-the-art FR-IQA models no matter for single distortion type or across whole database. Furthermore, we validated that the proposed metric performs better with shift-insensitive property compared with the CW-SSIM metric that is considered to be shift-insensitive IQA so far. Meanwhile, the proposed model is much simple than the CW-SSIM, which is efficient for applications., Comment: 10 pages, 7 figures

In video coding, it is expected that the encoder could adaptively select the encoding parameters (e.g., quantization parameter) to optimize the bit allocation to different sources under the given constraint. However, in hybrid video coding, the dependency between sources brings high complexity for the bit allocation optimization, especially in the block-level, and existing optimization methods mostly focus on frame-level bit allocation. In this paper, we propose a macroblock (MB) level bit allocation method based on the minimum maximum (MINMAX) criterion, which has acceptable encoding complexity for offline applications. An iterative-based algorithm, namely maximum distortion descend (MDD), is developed to reduce quality fluctuation among MBs within a frame, where the Structure SIMilarity (SSIM) index is used to measure the perceptual distortion of MBs. Our extensive experimental results on benchmark video sequences show that the proposed method can greatly enhance the encoding performance in terms of both bits saving and perceptual quality improvement., 11 pages, 17 figures

Among the various image quality assessment (IQA) tasks, blind IQA (BIQA) is particularly challenging due to the absence of knowledge about the reference image and distortion type. Features based on natural scene statistics (NSS) have been successfully used in BIQA, while the quality relevance of the feature plays an essential role to the quality prediction performance. Motivated by the fact that the early processing stage in human visual system aims to remove the signal redundancies for efficient visual coding, we propose a simple but very effective BIQA method by computing the statistics of self-similarity (SOS) in an image. Specifically, we calculate the inter-scale similarity and intra-scale similarity of the distorted image, extract the SOS features from these similarities, and learn a regression model to map the SOS features to the subjective quality score. Extensive experiments demonstrate very competitive quality prediction performance and generalization ability of the proposed SOS based BIQA method.

In the scatter correction for x-ray Cone Beam (CB) CT, the single-scan scheme with moving Beam Stop Array (BSA) offers reliable scatter measurement with low dose, and by using Projection Correlation based View Interpolation (PC-VI), the primary fluence shaded by the moving BSA (during scatter measurement) could be recovered with high accuracy. However, the moving BSA may increase the mechanical burden in real applications. For better practicability, in this paper we proposed a PC-VI based single-scan scheme with a ring-shaped stationary BSA, which serves as a virtual moving BSA during CB scan, so the shaded primary fluence by this stationary BSA can be also well recovered by PC-VI. The principle in designing the whole system is deduced and evaluated. The proposed scheme greatly enhances the practicability of the single-scan scatter correction scheme., 4 pages, Proceedings of the 11th International Meeting on Fully Three-dimensional Image Reconstruction in Radiology and Nuclear Medicine, Potsdam, Germany, 2011

Since Hounsfield's Nobel Prize winning breakthrough decades ago, X-ray CT has been widely applied in the clinical and preclinical applications - producing a huge number of tomographic gray-scale images. However, these images are often insufficient to distinguish crucial differences needed for diagnosis. They have poor soft tissue contrast due to inherent photon-count issues, involving high radiation dose. By physics, the X-ray spectrum is polychromatic, and it is now feasible to obtain multi-energy, spectral, or true-color, CT images. Such spectral images promise powerful new diagnostic information. The emerging Medipix technology promises energy-sensitive, high-resolution, accurate and rapid X-ray detection. In this paper, we will review the recent progress of Medipix-based spectral micro-CT with the emphasis on the results obtained by our team. It includes the state- of-the-art Medipix detector, the system and method of a commercial MARS (Medipix All Resolution System) spectral micro-CT, and the design and color diffusion of a hybrid spectral micro-CT.

This paper presents a comprehensive analysis on the security of the Yi-Tan-Siew chaotic cipher proposed in [IEEE TCAS-I 49(12):1826-1829 (2002)]. A differential chosen-plaintext attack and a differential chosen-ciphertext attack are suggested to break the sub-key K, under the assumption that the time stamp can be altered by the attacker, which is reasonable in such attacks. Also, some security Problems about the sub-keys $\alpha$ and $\beta$ are clarified, from both theoretical and experimental points of view. Further analysis shows that the security of this cipher is independent of the use of the chaotic tent map, once the sub-key $K$ is removed via the proposed suggested differential chosen-plaintext attack., Comment: 5 pages, 3 figures, IEEEtrans.cls v 1.6

This paper presents a general 3-D visualization method of cerebral vascular tree from Magnetic Resonance Angiography (MRA) images. First Gamma value and brightness are modified and each image is turned into binary ones. Then, vessels are extracted by the computation and then adaptive thresholding of region's areas. We stack the vessels to form 3-D volume data. The result is visualized in ANALYZE after linear interpolation using volume rendering. Satisfactory results to real angiographic data are obtained. It is helpful for analyzing and diagnosing the pathological changes of cerebral vessels.

Contour based techniques have proved to be an effective approach in object recognition. Active contour models (also called snakes), which optimize/minimize an energy function, have become popular for boundary detection. A snake is confused by a highly convex boundary. We present a novel adaptive algorithm to solve this problem. For every point in the initial position, the energy of its neighboring points is calculated by a greedy algorithm. If the target contour is not included in its neighbors, we can increase the radius of its neighbors and calculate the energy of all the points again until the target contour is included. The target contour can be obtained by iterating once. In addition, the convergent radius is increased. It can be applied to objects of high convexity. Comparative experiments indicate the validity of this method.

A novel D/sup 2/PCM compression technique has been developed in this paper for more effective lossless compression of DSA (digital subtraction angiography) image sequences, in which the correlation between subtraction images remains meaningful. By applying the lossless DPCM coder to the subtraction image, it constitutes a two-order differential coding scheme with the time variable. This proposed method is able to sufficiently reduce the temporal-space correlation. Higher compression ratio is achieved than previous techniques. The computational time and space also prove practicable and robust in real-life applications.

This paper proposes a new model for face detection based on statistics by applying the mosaic method. The model is insensitive to illumination and can adapt to the distortion and slope of face in some degrees. Moreover, there is an extra good feature of this model, i.e., an attention focusing capability. First, a distance measure, that is the Mahalanobis distance (MD), is presented, which is suitable for vector clustering. Next, the standard deviation of gradient norm and direction angle in each mosaic block are selected as the regional features in the model. The tested images are detected as human faces when both the MDs of the gradient norm and the direction angle to the collectivity of human faces are small. Finally, some experiments were made under different conditions and the results obtained show the model performed as expected.

The applications of digital chaotic maps in discrete-time chaotic cryptography and pseudo-random coding are widely studied recently. However, the statistical properties of digital chaotic maps are rather different from the continuous ones, which impedes the theoretical analyses of the digital chaotic ciphers and pseudo-random coding. This paper detailedly investigates the statistical properties of a class of digital piecewise linear chaotic map (PLCM), and rigorously proves some useful results. Based on the proved results, we further discuss some notable problems in chaotic cryptography and pseudo-random coding employing digital PLCM-s. Since the analytic methods proposed in this paper can essentially extended to a large number of PLCM-s, they will be valuable for the research on the performance of such maps in chaotic cryptography and pseudo-random coding.

Chaotic cryptology is widely investigated recently. This paper reviews the progress in this area and points out some existent problems in digital chaotic ciphers. As a comprehensive solution to these problems, a novel pseudo-random bit generator based on a couple of chaotic systems called CCS-PRBG is presented. Detailed theoretical analyses show that it has perfect cryptographic properties, and can be used to construct stream ciphers with higher security than other chaotic ciphers. Some experiments are made for confirmation. Finally, several examples of stream ciphers based on digital CCS-PRBG are given, and their security is discussed.

Registration is a problem of motion estimation (ME). The conventional ME methods often can not trade off between the estimation precision and computational complexity. This proposed method based on wavelet transform is a multi-scale block matching one, which takes the full advantage of the self-similarity that occurs after the wavelet transforming. The precision is from coarse to fine, which accelerates the matching procedure and reduces the computation consumption noticeably. This multi-scale registration algorithm has been proved practicable and useful in the real application.