Search

The accurate identification of the origin of Chinese medicinal materials is crucial for the orderly management of the market and clinical drug usage. In this study, a deep learning-based algorithm combined with machine vision was developed to automatically identify the origin of Angelica sinensis (A. sinensis) from eight areas including 1859 samples. The effects of different datasets, learning rates, solver algorithms, training epochs and batch sizes on the performance of the deep learning model were evaluated. The optimized hyperparameters of the model were the dataset 4, learning rate of 0.001, solver algorithm of rmsprop, training epochs of 6, and batch sizes of 20, which showed the highest accuracy in the training process. Compared to support vector machine (SVM), K-nearest neighbors (KNN) and decision tree, the deep learning-based algorithm could significantly improve the prediction performance and show better robustness and generalization performance. The deep learning-based model achieved the highest accuracy, precision, recall rate and F1_Score values, which were 99.55%, 99.41%, 99.49% and 99.44%, respectively. These results showed that deep learning combined with machine vision can effectively identify the origin of A. sinensis. [ABSTRACT FROM AUTHOR]

Deep learning models have been widely used in data-driven bridge structural damage diagnosis methods in recent years. However, these methods require training and test datasets to satisfy the same distribution, which is difficult to satisfy in practice. Domain adaptation transfer learning is an efficient method to solve this problem. Most of the current domain adaptation methods focus on close-set scenarios with the same classes in the source and target domains. However, in practical applications, new damage caused by long-term degradation often makes the target and source domains dissimilar in the class space. For such challenging open-set scenarios, existing domain adaptation methods will be powerless. To effectively solve the above problems, an adversarial auxiliary weighted subdomain adaptation algorithm is proposed for open-set scenarios. Adversarial learning is introduced to proposed an adversarial auxiliary weighting scheme to reflect the similarity of target samples with source classes. It effectively distinguishes unknown damage from known states. This paper further proposes a multi-channel multi-kernel weighted local maximum mean discrepancy metric (MCMK-WLMMD) to capture the fine-grained transferable information for conditional distribution alignment (sub-domain alignment). Extensive experiments on transfer tasks between three bridges verify the effectiveness of the algorithm in open-set scenarios. [ABSTRACT FROM AUTHOR]

The successful application of deep learning in bridge damage diagnosis relies on the assumption that the training and test data sets obey the same distribution. However, it is difficult to obtain labeled data of damage status for a bridge in using. Otherwise, it is difficult to apply a model trained with bridge A (source domain) to diagnose bridge B (target domain) because of the distribution discrepancy of data from different working environments or bridges. In response to these problems, motivated by transfer learning, a new bridge damage diagnosis method, namely, the multichannel domain adaptation deep transfer learning based method (MDADTL), is proposed in this paper. First, a CNN based multichannel multi‐scale feature extractor is introduced to extract features. Second, a multichannel domain adaptation module based on maximum mean discrepancy (MMD) is proposed for transfer learning, so that the learned features are domain‐invariant. Through the above process, MDADTL trained with labeled data obtained in the laboratory or the testing bridge is expected to diagnose other bridges with unlabeled data. Experiments prove the effectiveness and advancement of the proposed method. This exploration will promote the practical application of deep learning in bridge damage diagnosis. © 2022 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC. [ABSTRACT FROM AUTHOR]

Deep learning has been widely used in various fields because of its accuracy and efficiency. At present, the improvement of image semantic segmentation accuracy has become the area of most concern. In terms of increasing accuracy, improved semantic segmentation models have attracted more attention. In this paper, a hybrid model is proposed to solve the problems of edge splitting and small objects disappearing from complex scene images. The hybrid model consists of three parts: (1) an improved HED network, (2) an improved PSP-Net, (3) an AFF attention mechanism. Continuous edges can be obtained by combining the improved HED network with an improved PSP-Net. The AFF attention mechanism can improve the segmentation effect of small target objects by enhancing its response recognition ability for specific semantic scenes. The experiments were carried out on Cityspaces, SIFT Flow, NYU-V2 and CamVid datasets, and the experimental results show that the segmentation accuracy of our method is improved by 2% for small target objects, and by 3% for scenes with complex object edges. [ABSTRACT FROM AUTHOR]

Deep unsupervised learning extracts meaningful features from unlabeled images and simultaneously serves downstream tasks in computer vision. The basic process of deep clustering methods can include features learning and clustering assignment. To enhance the discriminative ability of the features and further improve the clustering performances, a new deep clustering method namely ACMEC (asymmetric convolutional denoising autoencoder with manifold spatial embedding clustering) is proposed. In this method, an asymmetric convolution denoising autoencoder is employed to extract visual features from images, and a manifold learning algorithm is used to obtain more distinctive features, followed by a Gaussian Mixture Model (GMM) is for clustering learning. The stability of feature space is guaranteed using separately training mechanism. In addition, reconstruction from noisy images enhances the robustness of feature networks. Experimental results on nine benchmark datasets demonstrate that the proposed ACMEC method can provide the better performances such as 0.979 clustering accuracy on the MNIST dataset and 0.668 on the fashion-MNIST dataset. ACMEC is a comparable competitor to the N2D (not too deep clustering) algorithm that is with 0.979 and 0.672 clustering accuracies respectively. Moreover, it is 16.1% higher than DEC algorithm on the fashion-MNIST dataset. [ABSTRACT FROM AUTHOR]

Deep learning based bridge damage diagnosis methods can successfully use labeled data to detect bridge damage. These successful applications usually need that the training samples (source domain) and test samples (target domain) obey the same probability distribution. However, it is difficult to acquire a large amount of labeled data with damage information from actual bridges. It is also difficult to apply a model trained with bridge A to diagnose bridge B because of the distribution discrepancy of data from different bridges or environments. Therefore, transferring a well-trained damage diagnosis model to another bridge with unlabeled data remains a major challenge. Motivated by transfer learning, this paper proposes a new intelligent damage diagnosis method for bridges, namely, sub-domain adaptive deep transfer learning network (SADTLN), to solve the feature generalization problem in different bridges. In our method, a multi-kernel local maximum mean discrepancy (MK-LMMD) based sub-domain adaptation module, including a domain classifier for aligning the global distribution and a sub-domain multi-layer adaptation for aligning local distribution, is proposed for transfer learning, so that the learned features are domain-invariant. Experiments prove the effectiveness and advancement of the proposed method. This exploration will promote the practical application of intelligent bridge structural damage diagnosis. [ABSTRACT FROM AUTHOR]

The autonomous and efficient learning of sewing gestures by robots will bring great convenience to the garment industry. To improve the accuracy of robots in detecting sewing gestures with high similarity, three detection models based on deep learning are proposed in the paper. First, in order to improve the detection accuracy and detection speed of sewing gestures under complex backgrounds, we added a dense connection layer to the low-resolution network layer of YOLO-V3 to enhance the transmission and reuse rate of image features. Secondly, a deeper ResNet50 residual network is introduced to replace the VGG16 basic network in the original SSD model. The feature pyramid structure is used to fuse high-level semantic features and low-level semantic features, which can improve the detection accuracy of small-sized sewing gestures. Finally, the parallel spatial-temporal dual-stream network separately extracts the temporal feature and the spatial feature of sewing gestures. The fusion of time feature and space feature improves the detection accuracy of the coherent sewing gesture. The results show that the suggested three models can effectively detect four sewing gestures with high similarity. Among them, the spatial-temporal two-stream convolutional neural network has the highest detection accuracy. The improved SSD model has faster detection speed than the improved YOLO-V3 model and other mainstream algorithms. [ABSTRACT FROM AUTHOR]

Occlusion is one of the key factors affecting the success rate of vision‐based fruit‐picking robots. It is important to accurately locate and grasp the occluded fruit in field applications, However, there is yet no universal and effective solution. In this paper, a high‐precision estimation method of spatial geometric features of occluded targets based on deep learning and multisource images is presented, enabling the selective harvest robot to envision the whole target fruit as if its occlusions do not exist. First, RGB, depth and infrared images are acquired. And pixel‐level matched RGB‐D‐I fusion images are obtained by image registration. Second, aiming at the problem of detecting the occluded tomatoes in the greenhouse, an extended Mask‐RCNN network is designed to extract the target tomato. The target segmentation accuracy is improved by 7.6%. Then, for partially occluded tomatoes, a shape and position restoration method is used to recover the obscured tomato. This algorithm can extract tomato radius and centroid coordinates directly from the restored depth image. The mean Intersection over Union is 0.895, and the centroid position error is 0.62 mm for the occluded rate under 25% and the illuminance between 1 and 12 KLux. And hereby a dual‐arm robotic harvesting system is improved to achieve a picking time of 11 s per fruit, an average gripping accuracy of 8.21 mm, and an average picking success rate of 73.04%. The proposed approach realizes a high‐fidelity geometrics reconstruction instead of mere image style restoration, which endows the robot with the ability to see through obstacles in the field scenes and improves its operational success rate in its result. [ABSTRACT FROM AUTHOR]

Single cellular lasers have recently attracted tremendous research due to their outstanding lasing characteristics for cell sensing and tracking. Thanks to enhanced light−cell interactions in Fabry–Pérot microcavities, transverse laser modes from cellular lasers are highly correlated to the spatial biophysical properties of cells. However, the huge complexity and randomness of laser modes set a critical challenge towards practical applications in cell analysis. In this study, deep learning is applied to unravel the complex laser modes generated from single‐cell lasers by establishing the correlation between laser modes and cellular physical properties. Primary cells extracted from rat brains and cell‐like droplets are investigated and trained through a convolutional neuron network based on laser mode images. Detailed simulations and experiments are conducted to study the effect of cell size on laser modes. Predictions of cell diameters with a sub‐micron accuracy are achieved with deep learning. Finally, the potential application of using deep‐learning‐enhanced laser modes for cell classification is demonstrated. Neuron and glial cells extracted from rat brains are classified through hyperspectral images of laser modes. The results demonstrate that deep learning has the potential to enable laser modes with biological significance and functions, offering new possibilities for biophotonic applications. [ABSTRACT FROM AUTHOR]

In recent years, intelligent structural damage diagnosis algorithms using machine learning have achieved much success. However, because of the fact that in real bridge applications, the working environment (load, temperature, and noise) is changing all the time, degradation of the performance of intelligent structural damage diagnosis methods is very serious. To address these problems, a novel bridge diagnosis algorithm based on deep learning is proposed. Our contributions include: First, we proposed an improved denoising auto‐encoder‐based deep neural networks, which is optimized by the gray relational analysis. It is able to automatically extract high‐level features from raw signals via a multi‐layer extraction to satisfy any damage diagnosis objective and thus does not need any time consuming denoising prepossessing. The model can achieve high accuracy under noisy environment. Second, the algorithm does not rely on any domain adaptation algorithm or require information of the target domain. It can achieve high accuracy when working environment is changed. Numerical simulations and experimental investigations on real bridges conducted to present the accuracy and efficiency of the proposed algorithm, comparing with other commonly machine learning‐based algorithms. The result shows it is deemed as an ideal and effective method for damage diagnosis of bridge structures. © 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC. [ABSTRACT FROM AUTHOR]

The Point Fractal Network (PF-Net) is a seminal work with capability of completing the missing regions of point clouds. However, the multi-resolution structure of PF-Net neglects effective feature fusion between each resolution, which makes the resulting shape lack local geometric details. To tackle this problem, we design a novel shape completion network named PRSCN. We first present Point Rank Sampling to rate and sample feature points more objectively through local outline form. In this way, the sampled points can facilitate the downstream tasks. Subsequently, considering the correlation between features from different scales, we design a Cross-Cascade Module to combine features hierarchically. Moreover, we propose Leap-type EdgeConv to enlarge the receptive field while maintain the kernel size unchanged. These improvements together make our CD error 3 % lower than that of state-of-the-art method on ShapeNet-part dataset. [ABSTRACT FROM AUTHOR]

Efficient and accurate distinction of histopathological subtype of lung cancer is quite critical for the individualized treatment. So far, artificial intelligence techniques have been developed, whose performance yet remained debatable on more heterogenous data, hindering their clinical deployment. Here, we propose an end-to-end, well-generalized and data-efficient weakly supervised deep learning-based method. The method, end-to-end feature pyramid deep multi-instance learning model (E2EFP-MIL), contains an iterative sampling module, a trainable feature pyramid module and a robust feature aggregation module. E2EFP-MIL uses end-to-end learning to extract generalized morphological features automatically and identify discriminative histomorphological patterns. This method is trained with 1007 whole slide images (WSIs) of lung cancer from TCGA, with AUCs of 0.95–0.97 in test sets. We validated E2EFP-MIL in 5 real-world external heterogenous cohorts including nearly 1600 WSIs from both United States and China with AUCs of 0.94–0.97, and found that 100–200 training images are enough to achieve an AUC of > 0.9. E2EFP-MIL overperforms multiple state-of-the-art MIL-based methods with high accuracy and low hardware requirements. Excellent and robust results prove generalizability and effectiveness of E2EFP-MIL in clinical practice. Our code is available at https://github.com/raycaohmu/E2EFP-MIL. [Display omitted] • A novel end-to-end and highly-generalized weakly supervised deep learning method to aid in lung cancer subtype diagnosis. • Patch sampling module and feature aggregation module to learn features more effectively and comprehensively. • Data efficient, 100 to 200 slides instead of thousands of ones is enough obtain test AUCs of > 0.9. • International, multicentre evaluation showing a promising generalizability on more heterogenous real-world data. • Results better than the other deep learning methods in the identification of lung cancer subtypes. [ABSTRACT FROM AUTHOR]

• A novel attention-guided deep learning method to aid in cervical cancer screening. • Attention module to mimic the way pathologist reading a cervical cytology image in clinical workflow. • Preparing a dataset of around 70,552 cervical cytology images for abnormal cervical cell detection. • Comparing the performance of our method with the other related deep learning methods as well as human pathologists. • Results better than the other deep learning methods and comparable to human pathologists with 10-year experience. [Display omitted] Early detection of abnormal cervical cells in cervical cancer screening increases the chances of timely treatment. But manual detection requires experienced pathologists and is time-consuming and error prone. Previously, some methods have been proposed for automated abnormal cervical cell detection, whose performance yet remained debatable. Here, we develop an attention feature pyramid network (AttFPN) for automatic abnormal cervical cell detection in cervical cytology images to assist pathologists to make a more accurate diagnosis. Our proposed method consists of two main components. First, an attention module mimicking the way pathologists reading a cervical cytology image. It learns what features to emphasize or suppress by refining extracted features effectively. Second, a multi-scale region-based feature fusion network guided by clinical knowledge to fuse the refined features for detecting abnormal cervical cells at different scales. The region proposals in the multi-scale network are designed according to the clinical knowledge about size and shape distribution of real abnormal cervical cells. Our method, trained and validated with 7030 annotated cervical cytology images, performs better than the state of art deep learning-based methods. The overall sensitivity, specificity, accuracy, and AUC of an independent testing dataset with 3970 cervical cytology images is 95.83%, 94.81%, 95.08% and 0.991, respectively, which is comparable to that of an experienced pathologist with 10 years of experience. Besides, we further validated our method on an external dataset with 110 cases and 35,013 images from a different organization, the case-level sensitivity, specificity, accuracy, and AUC is 91.30%, 90.62%, 90.91% and 0.934, respectively. Average diagnostic time of our method is 0.04s per image, which is much quicker than the average time of pathologists (14.83s per image). Thus, our AttFPN is effective and efficient in cervical cancer screening, and improvement of clinical workflows for the benefit of potential patients. Our code is available at https://github.com/cl2227619761/TCT_Detection. [ABSTRACT FROM AUTHOR]