Search

Stroke, particularly ischemic stroke, is a major cause of disability and one of the leading causes of adult mortality worldwide. Early and prompt management of stroke patients can reduce the severity of the disease. Doctors usually determine the severity of a stroke by focusing on the region of interest (ROI) in the MRI or CT scan images. An accurate and effective automatic image segmentation system can assist medical professionals as well as automatic detection and classification systems. Deep learning is the current advanced approach for dealing with machine learning and artificial intelligence. However, conventional deep learning requires a large amount of data for training, and the amount of labeled data in the medical field is limited. In this paper, we propose a few-shot learning strategy and integrate it with a base convolutional neural network model, which utilizes a self-attention mechanism to segment MRI for ischemic stroke. By combining the base model with self-attention, we can focus more on the ROI and disregard less important features. Additionally, the proposed system only selects slices with lesions and ignores unlesioned slices. This helps to improve efficiency and reduce the computational load by eliminating the need to tune unnecessary parameters. To achieve even better results, the system also combines two weighted images, FLAIR and DWI, as an early fusion process. Experiments have shown that this approach leads to higher performance compared to using the same system without fusion. The proposed system is evaluated using a publicly available dataset, ISLES 2015 SSIS, and compared with other state-of-the-art (SOTA) systems. It achieves a dice coefficient score of 0.68, which is significantly better than that of other SOTA systems. • An automatic ischemic stroke segmentation system using few-shot learning is proposed. • The proposed system selects only slices with lesions and excludes unlesioned slices. • We adopted prototypical networks with metric learning to achieve few-shot segmentation. • Two weighted MRIs, FLAIR, and DWI are fused in the few-shot learning paradigm. [ABSTRACT FROM AUTHOR]

Due to the influence of light absorption and scattering in the water medium, underwater images often appear with different degrees of blur, color distortion, and so on. The physical model-based method needs to establish image degradation processes and improve image quality by assuming and estimating complex parameters, which limits the generalization ability and practical application effect of these methods. To solve the above problems, this paper proposes an image enhancement method of non-uniform scattering in turbid underwater based on neural network. First, the high-frequency and low-frequency components of underwater images were obtained through the difference of gaussian (DOG) and bilateral filtering (BF), respectively. Then, for the high-frequency component, pulse coupled neural networks (PCNN) to improve the edge details of the image. For low-frequency components, it is carried out in two steps: an adaptive automatic white balance and polynomial regression (AAWBPR) algorithm is proposed, which can effectively eliminate the color bias under different color temperatures and improve the brightness of images; the adaptive color and contrast enhancement (ACCE) method is used to further enhance the color and contrast of images. Finally, Decom-Net and the improved U-Net method are used to reduce the errors caused by the non-uniform scattering effect, weaken the noise interference, and enhance the overall effect of images. Extensive experiments show that the proposed method can effectively improve the quality of turbid underwater non-uniform scattering images, enhance the contrast of images, reduce the complexity of parameters, and has good generalization performance and practical application effects. • A suppression algorithm of heterogeneous scattering in turbid water is proposed. • An adaptive automatic white balance and polynomial regression algorithm is proposed. • It reduces the image deviation, and improve the problem of complex model parameters. • It is applied to image segmentation, saliency detection, and keypoint detection. • This method can get better training results in small dataset. [ABSTRACT FROM AUTHOR]

• A novel model for improving identification of Diabetic Foot Ulcers. • A bottleneck attention module focused on channel-wise and spatial important features. • Experimental results show that the proposed model outperforms state-of-the-art models. Diabetic Foot Ulcer (DFU) is a severe complication of diabetes, and it may cause lower limb amputation. However, the manual diagnosis of DFU is a complicated and expensive process. The primary objective of this work is to design an efficient Convolutional Neural Network (CNN) approach to identify DFU. Therefore, a novel CNN-based approach (AESPNet) is proposed in this paper, where convolution layers are stacked together in a parallel fashion and with an intermediate attention module to perform DFU vs -normal skin classification. The AESPNet consists of 2 blocks, where varying-sized kernel convolution layers are connected in a parallel fashion for better local and global feature abstraction. A bottleneck attention module is associated after every concatenation operation in the network. The Stochastic Gradient Descent (SGD) (with momentum) optimizer with 1 e - 2 learning rate is used to train the proposed network on a privately accessed DFU dataset. The results of the proposed approach are compared with other standard CNN-based schemes, namely AlexNet, VGG16, DenseNet121, and InceptionV3. It has been found that the proposed AESPNet outperforms state-of-the-art schemes with a sensitivity score of 98.44% and 0.98 F1-Scores. [ABSTRACT FROM AUTHOR]

• A novel end-to-end network is proposed for fine-grained visual categorization. • The self-calibrated convolution is introduced to extract the salient features. • The gradual attention is proposed to find more local detailed features. • The inverse module is designed to boost the performance. Fine-grained visual categorization, which aims to identify the different subcategories of images within the same category, is a very challenging task due to the large intra-class differences and subtle inter-class variances. The existing methods mostly focus on the salient local regions and ignore other features which probably help to recognize the images more precisely. To address this issue, in this paper, we propose a novel end-to-end network composed of the self-calibrated convolution, gradual attention module and feature inverse module for fine-grained visual categorization. To extract the salient features, the self-calibrated convolution is exploited which can avoid the influence of irrelevant information and locate salient regions more accurately. In aiming to extract the discriminative features, we propose the gradual attention module which consists of alternate channel-spatial attention and hierarchical feature grouping. The gradual attention module can extract the subtle discriminative features gradually even when the semantic information of shallow stages is not rich. Moreover, we design the feature inverse module which forces the next stage of network to search for other different useful features by feature inverse. The gradual attention module combined with the feature inverse module is capable of finding more detailed regions and of benefit to improving classification performance. Finally, the stage features and fused features are jointly used for classification. The proposed method is evaluated on three classical fine-grained image datasets and compared with a number of state-of-the-art methods. Our method achieves 89.5 % , 95.2 % and 93.9 % accuracies on CUB-200–2011, Stanford Cars and FGVC-Aircraft datasets respectively. The experimental results demonstrate the effectiveness and superiority of the proposed method. [ABSTRACT FROM AUTHOR]

• A novel joint-nullspace is carfully designed instead of feature space. • An adaptive loss function is developed to train the bounding box estimation. • Extensive experiments have been conducted to evaluate NullSpaceRDAR on six benchmarks. Recently, discriminative-based and Siamese-based trackers have achieved outstanding performance on most tracking benchmarks. However, these trackers use the pre-trained backbone networks that have been mainly designed for classification to extract target-specific features without taking into consideration the visual object tracking task. In this paper, we propose NullSpaceRDAR, a novel tracker that learns a robust target-specific feature representation specifically designed for object tracking. This feature representation is learned by projecting the traditional backbone feature space onto a novel discriminative nullspace that is used to regularize the backbone loss function. We refer to the discriminative nullspace herein as joint-nullspace. The same target features (i.e., target-specific) in the proposed joint-nullspace are collapsed into a single point, and different target-specific features are collapsed into different points. Consequently, the joint-nullspace forces the network to be sensitive to the object's variations from the same class (i.e., intra-class variations). Moreover, a modified adaptive loss function is developed for bounding box estimation to select the most suitable loss function from a super set family of loss functions based on the training data. This makes NullSpaceRDAR more robust to different challenges such as occlusions and background clutter. Extensive experiments have been conducted on six benchmarks to evaluate NullSpaceRDAR: OTB100, VOT variations (VOT2018, VOT2019, and VOT2020), LaSOT, TrackingNet, UAV123, and GOT10k. The results show that NullSpaceRDAR outperforms the state-of-the-art trackers. [ABSTRACT FROM AUTHOR]

Saliency detection models using light field data as input have not been thoroughly explored. Existing deep saliency models usually treat multi-focus images as independent information and extract their features separately, which may be cumbersome and over-rely on well-designed network structure. Besides, they do not fully explore the cross-modal complementarity and cross-level continuity of information, and rarely consider edge cues. Based on the above observations, in this paper, we investigate a novel Dual Guidance Enhanced Network (DGENet), which considers both spatial content and explicit boundary cues. Specifically, DGENet contains two key modules, i.e., the recurrent global-guided focus module (RGFM) and the boundary-guided semantic accumulation module (BSAM). These two modules are composed of multiple units, and the units in each module are not independent of each other. RGFM is used to distill out effective squeezed information of focal slices and RGB images between different levels. The learned global context features guide the network to focus on the salient region via a progressive reverse attention-driven strategy. Furthermore, BSAM introduces salient edge features to guide the accumulation of salient object features to generate salient maps with sharp boundaries. Extensive experiments on three challenging light field datasets demonstrate that our DGENet is superior to cutting-edge 2D, 3D and 4D methods. • A Dual Guidance Enhanced Network for light field saliency detection is proposed. • The global map as the initial guidance is obtained through global generation module. • Recurrent global-guided focus module integrating reverse attention-driven strategy. • Boundary-guided semantic accumulation module explicitly introduces salient edge cues. • Experimental results demonstrate the effectiveness and superiority of the method. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel framework to segment 3D human motion capture data into distinct behaviors. First, in preprocessing, we build a normalized pose space by eliminating translation and orientation from the 3D poses. We then transform these normalized 3D poses into 2D RGB images, and as a result, we simplify the task of motion segmentation as image classification and recognition. Furthermore, we identify the most significant joints of the skeleton that contribute substantially to executing a motion and get benefits from them by assigning them more weights. The weight allocation to the specific joint has been done purely based on its deviation capability. Finally, each motion is encoded into compact visual representation by exploiting RGB images with weighted joints. We adopt a transfer learning approach to extract a fixed-size feature vector using off-the-shelf deep Convolutional Neural Network (CNN), Alexnet, after fine-tuning. We develop a Kd-tree on these highly descriptive feature vectors to retrieve the nearest neighbors. Based on a similarity measure, we classify the motion segments and ultimately place the cuts on the ongoing motion sequences. We perform extensive experiments to evaluate our proposed approach on popular Motion Capture (MoCap) datasets, CMU and HDM05. Our approach almost outperforms all other state-of-the-art methods, and the results highlight the capabilities of our proposed scheme for effective segmentation. • We propose a novel data-driven approach for the segmentation of motion capture data. • We encode x, y, and z coordinates of a joint into an RGB image pixel. In this way, we develop an image that represents the complete motion. • We adopt a transfer learning approach and use a pre-trained deep CNN architecture, "Alexnet" to extract the fixed-size features. • We build Kd-tree on these fixed-size features to search and retrieve nearest neighbors from the input query frames. • Based on the retrieved nearest neighbors, the actions are classified, and the segmentation is performed. [ABSTRACT FROM AUTHOR]

Efficient semantic segmentation of large-scale three-dimensional (3D) point clouds is an essential approach for intelligent robots to perceive the surrounding environment. However, due to the expensive sampling process or time-consuming pre/post-processing steps, most of the current solutions are inefficient or limited in scale. In this paper, we propose a novel framework, ThickSeg, to efficiently assign semantic labels for large-scale point clouds. ThickSeg contains three main steps: Firstly, it projects raw point clouds onto a multi-layer image with a random-hit strategy to efficiently preserve more local geometric features. Secondly, the projected multi-layer image is fed into a Self-Sorting 3D Convolutional Neural Network (SS-3DCNN) to predict grid-wise semantics and subsequently project them back to their corresponding 3D points. Finally, the labels of occluded points are determined by an iterative and accumulative post-processing mechanism, avoiding time-consuming explicit 3D neighborhood searching. We validate our approach on two well-known public benchmarks (SemanticKITTI and KITTI), where ThickSeg gets state-of-the-art results and more efficient than previous methods. Our detailed ablation study shows how each component contributes to the final performance. • Introduce a novel framework for efficient semantic segmentation of point clouds. • Project raw point clouds into a multi-layer image to leave more details. • Design a layer adaptive neural network for pixel-wise semantics. • Determine semantics of occluded points in an iterative and accumulative manner. [ABSTRACT FROM AUTHOR]

How to effectively utilize and fuse deep features has become a critical point for salient object detection. Most existing methods usually adopt the convolutional features based on U-shape structures and fuse multi-scale convolutional features without fully considering the different characteristics between high-level features and low-level features. Furthermore, existing salient object detection methods rarely consider the role of pooling in convolutional neural networks. Moreover, there is still much room to improve the detection performance for objects in complex scenes. To address the problems mentioned above, we propose a pooling-based feature pyramid (PFP) network to boost salient object detection performance in this paper. First, we design two U-shaped feature pyramid modules to capture rich semantic information from high-level features and to obtain clear saliency boundaries from low-level features respectively. Second, a pyramid pooling refinement module is designed to utilize the pooling to capture more semantic information. Third, a universal channel-wise attention (UCA) module is designed to select effective high-level features of multi-scale and multi-receptive-field for rich semantic information, even in complex scenes. Finally, we fuse the selected high-level features and low-level features together, followed by an edge preservation loss to obtain accurate boundary location. Extensive experiments are conducted on five datasets and the experimental results indicate that our proposed method has the ability to get better salient object detection performance compared to the state-of-the-art methods. • A pooling-based feature pyramid network for salient object detection is proposed; • Two U-shaped feature pyramid modules are designed; • Pyramid pooling refinement module is designed to capture rich semantic information; • Universal channel-wise attention module is designed to select high-level features; • Experiments demonstrate the superiority of PFP for salient object detection. [ABSTRACT FROM AUTHOR]

Periocular region is considered as an important biometric trait due to its ease of collectability and high acceptability in our society. Recent advancements in surveillance applications require infra-red (IR) sensing equipments to be deployed in order to capture the activities occurring in low-light conditions. This gives rise to the problem of matching periocular images in heterogeneous environments as it is difficult to avail large enrollment datasets in IR modality within a short span of time. Although a number of approaches have studied cross-spectral matching of periocular images where a probe IR image is matched against enrolled dataset images in visible (VIS) domain and vice versa, significant amount of challenges still exists for such matching. In this paper, we propose an attention-based twin deep convolutional neural network (CNN) with shared parameters in order to match the periocular images in heterogeneous modality. We introduce a novel variance-guided objective function in conjunction with the attention module in order to guide the network to focus more into the relevant regions of the periocular images. The weights of the twin model based on the new objective function are learned so as to reduce the intra-class variance and to increase the inter-class variance of the cross-spectral image pairs. Ablation studies and experimental results on three publicly available cross-spectral periocular datasets containing images from VIS, near-infrared (NIR), and night vision domains show that the proposed deep network achieves the state-of-the-art recognition performances on all three datasets. • Cross-spectral matching reduces the gap amid periocular images in different spectra. • Attention module focuses on the significant regions of the heterogeneous input. • Variance-guided objective function accounts for both probability and feature space. • Intermediate layer visualizations and ablation studies show efficacy of VGACNet. [ABSTRACT FROM AUTHOR]

Deep learning has been applied to fashion landmark detection in recent years, and great progress has been made. However, the detection of hard keypoints, such as those which are occluded or invisible, remains challenging and must be addressed. To tackle this problem, in the feature exaction level a novel Multi-Depth Dilated (MDD) block which is composed of different numbers of dilated convolutions in parallel and a Multi-Depth Dilated Network (MDDNet) constructed by MDD blocks are proposed in this paper, and in the training level a network training method of Batch-level Online Hard Keypoint Mining (B-OHKM) is proposed. During the training of network, each clothing keypoint is one-to-one corresponding to the related loss value calculated at that keypoint. The greater the loss of the keypoint, the more difficult it is for the network to detect that keypoint. In that way, hard keypoints can be effectively mined, so that the network can be trained in a targeted manner to improve the performance of hard keypoints. The results of experiments on two large-scale fashion benchmark datasets demonstrate that the proposed MDDNet that uses the MDD block and B-OHKM method achieves state-of-the-art results. • A novel Multi-Depth Dilated (MDD)block that can efficiently extract different levels of large-scale context information, which is beneficial for the inference of hard keypoints, is proposed. • The Batch-level Online Hard Keypoints Mining(B-OHKM)method is proposed for training to further improve the effectiveness of hard keypoints detection. • It is demonstrated that anetwork (MDDNet) that uses the MDD block and B-OHKM training method obtains significant improvements over state-of-the-artmethodsonstandardbenchmarksforfashionlandmarkdetection. [ABSTRACT FROM AUTHOR]