Salient object detection (SOD) is a challenging task in computer vision. Current SOD approaches have made significant progress, but they fail in challenging scenarios. This paper categorizes the existing challenges in SOD into four groups: images ...

• Proposing the DASOD method for salient object detection (SOD) in challenging images.
• Identifying existing challenges and addressing them via camouflaged object detection.
• SOD in images with complex background, low contrast, ...

Unintentional human falls, especially in seniors, lead to serious injuries, fatalities, and reduced standard of life. Vision-based fall detection methods have demonstrated their usefulness in timely fall response, helping to lessen such injuries. ...

• A human fall detection system using a 4-stream 3D CNN is proposed.
• The human body is segmented from the video clip using ResNet18.
• The incident is divided into four sequences: standing, falling, fallen, and others.
• Frames of ...

The combination of visible and thermal images has been proven to be effective in improving accuracy for crowd counting in illumination-unconstrained scenes. However, the challenging problem of misalignment in RGB-T image pairs has not been ...

Display Omitted

• Explore the cross-modal misalignment issue for RGB-Thermal crowd counting task.
• Design a spatial semantic joint alignment block for misaligned cross-modal counting.
• Raise a low-frequency feature attention filtering block for the ...

With the rising popularity of omnidirectional images (ODIs) in virtual reality applications, the need for specialized image quality assessment (IQA) methods becomes increasingly critical. Traditional IQA approaches, designed for rectilinear ...

• Introduces LGT360IQ, a dual-branch framework for 360° image quality assessment (IQA).
• Mimics top-down and bottom-up visual attention for omnidirectional images (ODIs).
• Employs Transformers for local and global processing to ...

Change detection using remote sensing images captured by small unmanned aerial vehicles (small UAVs) finds wide applications across various fields. However, there is a challenge when dealing with images captured at the same location by small UAVs ...

• A novel change detection model for images with viewpoint differences is proposed.
• OFACD simultaneously aligns images and detects changes toenhanceperformance.
• Created two change detection datasets with viewpoint differences.
• ...

Nighttime image semantic segmentation is challenging due to low-light and diverse lighting conditions. A straightforward solution is to first enhance nighttime scene images to resemble daytime scene before performing segmentation. This kind of ...

• Developed a Retinex-based network RNightSeg to enhance nighttime image segmentation.
• Introduced illumination-independent reflectance component for accurate segmentation.
• Method improves performance across diverse nighttime ...

This study aims to enhance the robustness and generalization capability of a deep learning transformer model used for segmenting skin carcinomas and tissues through the introduction of incremental learning. Deep learning AI models demonstrate ...

• A novel 2-dimensional hybrid incremenatl learning framework for multi-class segmentation of skin tissues and non-melanoma skin cancer
• Combined knowledge distillation and the replay approach to address catastrophic forgetting
• This ...

Breast cancer continues to be a major global health challenge, necessitating reliable diagnostic methods for early detection and improved patient outcomes. This study introduces a novel ensemble fuzzy model for predictive breast cancer diagnosis, ...

• A new approach integrates deep-learning classifiers with fuzzy logic for improved decision-making.
• The ensemble includes Inception-V4, Inception-ResNet, and Inception V3/V4 + BN, enhancing accuracy.
• Fuzzy logic allows adaptive ...

Dynamic textures (DT) are typically 3D videos of physical processes showing statistical regularity but have indeterminate spatial and temporal extent. Existing DT recognition methods usually neglect the global spatio-temporal relationships of DT ...

• A novel SMTCNN architecture considering the global spatio-temporal relationships of dynamic textures is proposed for 3D dynamic texture recognition.
• We present a new contextual module transforming 2D CNN features to 1D vectors for 3D ...

3D image registration is now common in many medical domains. Multimodal registration implies the use of different imaging modalities, which results in lower accuracy compared to monomodal registration. The aim of this study ...

Display Omitted

• We translated a multimodal CBCT-MR/CT registration into a sCT/CT registration.
• The unsupervised synthesis method was based on a cGAN using novel perceptual loss.
• The best registration accuracy was obtained via synthetic image ...

Although existing facial expression recognition (FER) methods have achieved great success, their performance degrades significantly under noisy labels caused by low-quality images, ambiguous expressions, and subjective and incorrect labeling. ...

• An effective DC-Net is presented to learn robust representations to suppress noisy samples.
• A novel Class Activation Mapping Attention Consistency is proposed to make model focus on partially important features.
• A novel Class ...

Crowd counting is essential for video surveillance and public safety. The performance of counting models has been greatly improved with the rapid development of Convolution neural networks (CNN), while it still suffers interference from complex ...

• The insufficient scale hierarchy imposes some limitations on the performance.
• The network is inclined to confuse dense regions and background regions.
• Euclidean distance is strict pixel-level alignment will overfit the density map ...

Recently, the enhanced abilities of diffusion models have led to more realistic inpainting results, which raises the potential for criminal activity through image forgery. In this study, we explore the detection of inpainted images generated by a ...

Display Omitted

• We demonstrate the localization of inpainting images generated by diffusion models.
• We introduce a new image forensics dataset based on diffusion modelbased inpainting.
• We propose an inter-patch relationship through the learnable ...

This paper proposed a novel pretext task to address the skeleton-based video representation for self-supervised action recognition tasks. Instead of exploiting only the whole body, various levels of the skeleton structure (e.g., upper body, lower ...

This paper focuses on semi-supervised video object segmentation (VOS). Recently, several Space‐Time Memory based networks have effectively improved the performance of VOS. However, most methods predict the target object mask forwardly, which ...

• Propose a bidirectional strategy to alleviate the error propagation problem.
• Propose a multi-scale module refine the prediction of object boundaries.
• Non-local modules combined with channel attention.
• Experimental results show ...

Active domain adaptation aims to enhance model adaptation performance by annotating a limited number of informative unlabeled target data. Traditional active learning strategies for semantic segmentation often neglect the presence of domain ...

• We propose a novel active domain adaptation framework for semantic segmentation.
• The domainness score is designed to evaluate target samples' representativeness.
• The prediction uncertainty is utilized to evaluate target samples' ...

In the rapidly evolving domain of visual tracking, the accurate identification and tracking of small objects pose significant challenges due to their minimal pixel presence and detail deficiency. In the field of small object tracking, utilizing ...

• We present a small object tracker using reversible networks to preserve detailed information.
• We integrate two reversible networks: one for capturing small object details and another for enhancing distinguishability in complex ...

Recently, U-shaped neural networks have gained widespread application in remote sensing image dehazing and achieved promising performance. However, most of the existing U-shaped dehazing networks neglect the global and local information ...

• We propose a method (HSMD-Net) for haze removal in remote sensing images.
• A new reconstruction module and a new information interaction module.
• The method can produce reliable results under various concentrations of haze.
• ...

The visible-infrared person re-identification (VI-ReID) task is challenging in image retrievals because of the modality gaps between visible and infrared images. Different from the most existing methods which either strive to capture modality ...

• Establish an interactive relation between modality-shared features and modality-specific features.
• A modality interactive attention module is introduced to narrow down the modal gap.
• Two feature extraction strategies have been ...