Saliency detection is a fundamental yet challenging task in computer vision, aiming at highlighting the most visually distinctive objects an image. We propose novel recurrent residual refinement network (R^3Net) equipped with blocks (RRBs) to more accurately detect salient regions of input Our RRBs learn between intermediate saliency prediction and ground truth by alternatively leveraging low-level integrated features high-level fully convolutional (FCN). While are capable capturing details,...

This paper presents a novel unsupervised domain adaptation framework, called Synergistic Image and Feature Adaptation (SIFA), to effectively tackle the problem of shift. Domain has become an important hot topic in recent studies on deep learning, aiming recover performance degradation when applying neural networks new testing domains. Our proposed SIFA is elegant learning diagram which synergistic fusion adaptations from both image feature perspectives. In particular, we simultaneously...

Convolution on 3D point clouds that generalized from 2D grid-like domains is widely researched yet far perfect. The standard convolution characterises feature correspondences indistinguishably among points, presenting an intrinsic limitation of poor distinctive learning. In this paper, we propose Adaptive Graph (AdaptConv) which generates adaptive kernels for points according to their dynamically learned features. Compared with using a fixed/isotropic kernel, AdaptConv improves the...

Can you find me? By simulating how humans to discover the so-called 'perfectly'-camouflaged object, we present a novel boundary-guided separated attention network (call BSA-Net). Beyond existing camouflaged object detection (COD) wisdom, BSA-Net utilizes two-stream modules highlight separator (or say object's boundary) between an image's background and foreground: reverse stream helps erase interior focus on background, while normal recovers thus pay more foreground; both streams are...

Skin lesion segmentation from dermoscopy images is of great significance in the quantitative analysis skin cancers, which yet challenging even for dermatologists due to inherent issues, i.e., considerable size, shape and color variation, ambiguous boundaries. Recent vision transformers have shown promising performance handling variation through global context modeling. Still, they not thoroughly solved problem boundaries as ignore complementary usage boundary knowledge contexts. In this...

Convolution on 3D point clouds is widely researched yet far from perfect in geometric deep learning. The traditional wisdom of convolution characterises feature correspondences indistinguishably among points, arising an intrinsic limitation poor distinctive In this article, we propose Adaptive Graph (AGConv) for wide applications cloud analysis. AGConv generates adaptive kernels points according to their dynamically learned features. Compared with the solution using fixed/isotropic kernels,...

While the wisdom of training an image dehazing model on synthetic hazy data can alleviate difficulty collecting real-world hazy/clean pairs, it brings well-known domain shift problem. From a different yet new perspective, this paper explores contrastive learning with adversarial effort to leverage unpaired and clean images, thus alleviating problem enhancing network's generalization ability in scenarios. We propose effective unsupervised paradigm for dehazing, dubbed UCL-Dehaze. Unpaired...

Grading laryngeal squamous cell carcinoma (LSCC) based on histopathological images is a clinically significant yet challenging task. However, more low-effect background semantic information appeared in the feature maps, channels, and class activation which caused serious impact accuracy interpretability of LSCC grading. While traditional transformer block makes extensive use parameter attention, model overlearns information, resulting ineffectively reducing proportion semantics. Therefore,...

A role for autophagy, a conserved cellular response to stress, has recently been demonstrated in human cancers. Aberrant expression of Beclin-1, an important autophagic gene, reported various In the present study, we investigated significance and relationship between Beclin-1 cell proliferation, apoptosis, microvessel density (MVD) clinical pathological changes or prognosis hepatocellular carcinoma (HCC).A total 103 primary HCC patients were involved study. Expression PCNA, NET-1, Bcl-2,...

Adaptive person re-identification (adaptive ReID) targets at transferring learned knowledge from the labeled source domain to unlabeled target domain. Pseudo-label-based methods that alternatively generate pseudo labels and optimize training model have demonstrated great effectiveness in this field. However, generated are inaccurate cannot reflect true semantic meaning of samples. We consider such inaccuracy stems both lagged update as well simple criterion employed clustering method. To...

The tumor grading of laryngeal cancer pathological images needs to be accurate and interpretable. deep learning model based on the attention mechanism-integrated convolution (AMC) block has good inductive bias capability but poor interpretability, whereas vision transformer (ViT) interpretability weak ability. Therefore, we propose an end-to-end ViT-AMC network (ViT-AMCNet) with adaptive fusion multiobjective optimization that integrates fuses ViT AMC blocks. However, existing methods often...

We study the semi-supervised learning problem, using a few labeled data and large amount of unlabeled to train network, by developing cross-patch dense contrastive framework, segment cellular nuclei in histopathologic images. This task is motivated expensive burden on collecting for image segmentation tasks. The key idea our method align features teacher student networks, sampled from cross-image both patch- pixel-levels, enforcing intra-class compactness inter-class separability that as we...

Automatic polyp segmentation from colonoscopy videos is a prerequisite for the development of computer-assisted colon cancer examination and diagnosis system. However, it remains very challenging task owing to large variation polyps, low contrast between polyps background, blurring boundaries polyps. More importantly, real-time performance necessity this task, as anticipated that segmented results can be immediately presented doctor during intervention his/her prompt decision action. It...

One critical challenge in 6D object pose estimation from a single RGBD image is efficient integration of two different modalities, i.e., color and depth. In this work, we tackle problem by novel Deep Fusion Transformer (DFTr) block that can aggregate cross-modality features for improving estimation. Unlike existing fusion methods, the proposed DFTr better model semantic correlation leveraging their similarity, such globally enhanced modalities be integrated improved information extraction....

How will you repair a physical object with some missings? You may imagine its original shape from previously captured images, recover overall (global) but coarse first, and then refine local details. We are motivated to imitate the procedure address point cloud completion. To this end, we propose cross-modal shape-transfer dual-refinement network (termed CSDN), coarse-to-fine paradigm images of full-cycle participation, for quality CSDN mainly consists "shape fusion" "dual-refinement"...

Surgical instrument segmentation is fundamentally important for facilitating cognitive intelligence in robot-assisted surgery. Although existing methods have achieved accurate results, they simultaneously generate masks of all instruments, which lack the capability to specify a target object and allow an interactive experience. This paper focuses on novel essential task robotic surgery, i.e., Referring Video Instrument Segmentation (RSVIS), aims automatically identify segment surgical...

Multimodal neuroimaging provides complementary information critical for accurate early diagnosis of Alzheimer's disease (AD). However, the inherent variability between multimodal neuroimages hinders effective fusion features. Moreover, achieving reliable and interpretable diagnoses in field remains challenging. To address them, we propose a novel network based on multi-fusion disease-induced learning (MDL-Net) to enhance AD by efficiently fusing data. Specifically, MDL-Net proposes joint...

Toll-like receptor 3 (TLR3) plays a key role in innate immunity. In the present study, we analyzed tissues of patients with human hepatocellular carcinoma (HCC) to determine significance relationship between TLR3 expression and cell proliferation, apoptosis, hepatitis B virus infections, angiogenesis prognosis.We collected paraffin-embedded from 85 HCC who had complete histories were followed for >5 years. The intracellular localization downstream proteins (TRIF, NF-κB, IRF3) detected using...

We propose a novel convolutional neural network (ConvNet) equipped with two new semantic calibration and refinement approaches for automatic polyp segmentation from colonoscopy videos. While ConvNets set state-of-the-are performance this task, it is still difficult to achieve satisfactory results in real-time manner, which necessity clinical practice. The main obstacle the huge gap between high-level features low-level features, making take full advantage of complementary information...

Current representation learning methods for whole slide image (WSI) with pyramidal resolutions are inherently homogeneous and flat, which cannot fully exploit the multiscale heterogeneous diagnostic information of different structures comprehensive analysis. This paper presents a novel graph neural network-based multiple instance framework (i.e., H^2-MIL) to learn hierarchical from WSI A “resolution” attribute is constructed explicitly model feature spatial-scaling relationship...

Camouflaged objects share very similar colors but have different semantics with the surroundings. Cognitive scientists observe that both global contour (i.e., boundary) and local pattern texture) of camouflaged are key cues to help humans find them successfully. Inspired by cognitive scientist's observation, we propose a novel boundary-and-texture enhancement network (FindNet) for object detection (COD) from single images. Different most existing COD methods, FindNet embeds information into...

We present a novel deep network (namely BUSSeg) equipped with both within- and cross-image long-range dependency modeling for automated lesions segmentation from breast ultrasound images, which is quite daunting task due to (1) the large variation of lesions, (2) ambiguous lesion boundaries, (3) existence speckle noise artifacts in images. Our work motivated by fact that most existing methods only focus on within-image dependencies while neglecting dependencies, are essential this under...

Cancer survival prediction requires exploiting related multimodal information (e.g., pathological, clinical and genomic features, etc.) it is even more challenging in practices due to the incompleteness of patient's data. Furthermore, existing methods lack sufficient intra- inter-modal interactions, suffer from significant performance degradation caused by missing modalities. This manuscript proposes a novel hybrid graph convolutional network, entitled HGCN, which equipped with an online...

Current point cloud denoising (PCD) models optimize single networks, trying to make their parameters adaptive each in a large pool of clouds. Such network paradigm neglects that different points are often corrupted by levels noise and they may convey geometric structures. Thus, the intricacy both geometry poses side effects including remnant noise, wrongly-smoothed edges, distorted shape after denoising. We propose <italic xmlns:mml="http://www.w3.org/1998/Math/MathML"...

As a novel similarity measure that is defined as the expectation of kernel function between two random variables, correntropy has been successfully applied in robust machine learning and signal processing to combat large outliers. The usually zero-mean Gaussian kernel. In recent work, concept mixture (MC) was proposed improve performance, where kernel, namely, linear combination several kernels with different widths. both MC, center is, however, always located at zero. present further we...