24 December 2024 in IEEE Transactions on Multimedia
Camouflaged object detection (COD) aims to segment targeted objects that have similar colors, textures, or shapes to their background environment. Due to the limited ability in distinguishing highly similar patterns, existing COD methods usually produce inaccurate predictions, especially around the boundary areas, when coping with complex scenes. This paper proposes a Progressive Region-to-Boundary Exploration Network (PRBE-Net) to accurately detect camouflaged objects. PRBE-Net follows an encoder-decoder framework and includes three key modules. Specifically, firstly, both high-level and low-level features of the encoder are integrated by a region and boundary exploration module to explore their complementary information for extracting the object's coarse region and fine boundary cues simultaneously. Secondly, taking the region cues as the guidance information, a Region Enhancement (RE) module is used to adaptively localize and enhance the region information at each layer of the encoder. Subsequently, considering that camouflaged objects usually have blurry boundaries, a Boundary Refinement (BR) decoder is used after the RE module to better detect the boundary areas with the assistance of boundary cues. Through top-down deep supervision, PRBE-Net can progressively refine the prediction. Extensive experiments on four datasets indicate that our PRBE-Net achieves superior results over 21 state-of-the-art COD methods. Additionally, it also shows good results on polyp segmentation, a COD-related task in the medical field.
17 February 2025 in IEEE Transactions on Circuits and Systems for Video Technology
Colorectal polyp segmentation in endoscopic images is very important for the prevention and treatment of colorectal cancer. Because of the high similarity between polyps and their surrounding tissues, most deep neural network (DNN) based methods often struggle with blurry boundaries and result in inaccurate segmentation. In this paper, we propose a Boundary-guided Feature-aligned Network (BFNet) for polyp segmentation by taking a boundary prediction task as an auxiliary. Firstly, BFNet aggregates multi-layer features extracted from the backbone to mine boundary cues. Secondly, a flexible feature aggregation (FFA) module is used at each layer to adaptively fuse cross-layer features for coarse polyp localization. In the FFA module, considering the spatial misalignment between features at different layers, the feature of the high layer is aligned to and fused with that of the current layer using the deformable convolution and flexible merge block. After that, a boundary-guided feature enhancement (BFE) module is applied to refine the localization at boundary areas. In the BFE module, the boundary information is extracted and highlighted in both channel and spatial dimensions using the attention mechanisms with the assistance of boundary cues. By applying deep supervision to the BFE modules, BFNet can produce accurate polyp segmentation. Experimental results show that our BFNet outperforms 14 state-of-the-art DNN-based polyp segmentation methods on both in-domain and out-of-domain tests.
2 September 2024 in IEEE Transactions on Broadcasting
Face retouching involves using digital techniques to alter an individual’s appearance, commonly using in social media. However, excessively retouched face (RF) images can lead to issues such as unrealistic beauty standards and psychological stress. Therefore, it is crucial to develop a reliable quality assessment method for RF images. In this paper, we propose a novel network named DIRF-IQA for RF image quality assessment (IQA). DIRF-IQA mainly includes a parameter-shared image encoder, a text encoder, and three key components, namely the Differential Feature Attention Module (DFAM), the Text-image Interaction Module (TIM), and the Multi-scale Feature Fusion Module (MFFM). Specifically, the DFAM captures both local and global differences between original and retouched images by processing multi-scale features and utilizing cross-attention and self-attention blocks for differential perception. In the TIM, textual prompts summarizing retouching operations are encoded by a text encoder and integrated with differential features extracted by the DFAM to enhance the understanding of distortions in RF images. The MFFM then fuses these text-enhanced features across different layers and combines them with the global differential feature to predict the quality of the retouched images. We conduct extensive experiments on two RF IQA databases and the results demonstrate the superiority of DIDF-IQA compared to 12 state-of-the-art full-reference IQA methods in evaluating RF images.
11 October 2024 in IEEE Sensors Journal
To improve the measurement accuracy of Raman distributed temperature sensors (RDTS) in a simplified and efficient manner has always been of paramount importance. Based on the characteristics of Raman backscattered light, this paper proposes a new demodulation optimization method based on multiple compensations and linear fitting, which effectively improves the measurement accuracy of RDTS. Merely describing the relationship between temperature and Raman light intensity ratio through a linearly fitted relational expression does not result in high measurement accuracy. Moreover, when the system operates for a long time, the sensing curve may shift as the reference temperature rises due to the structure of the calibration module. To address these issues, we have added two techniques based on linear fitting: proportional coefficients adjustment algorithm and reference temperature offset compensation. Firstly, we use the proportional coefficients adjustment algorithm to solve the problem of insufficient temperature response in the linear model. Then, we address the issue of sensing curve shifts through reference temperature offset compensation model. Finally, this method has achieved great results, improving the measurement accuracy of the sensing system from ±7.8°C to ±0.8°C within a 2.4km measurement range.
1 April 2025 in IEEE Sensors Journal
Accurate polyp segmentation in colonoscopy images is crucial for early diagnosis and timely treatment of colorectal cancer. Recently, deep learning-based methods exhibit distinct advantages in polyp segmentation. However, most existing methods often obtain ordinary performance due to limited ability in extracting context information and foreground cues. To overcome these drawbacks, we propose a Context-Aware Deep Network (CADNet) with cross-layer feature fusion for polyp segmentation in this paper. Specifically, CADNet is designed with an encoder-decoder framework. In the encoder stage, a group guidance context module is proposed for the top-three highest layers to make the network focus more on target regions by aggregating multi-scale context information with the prior knowledge from the adjacent high layer. In the decoder stage, a cross-layer feature fusion module is proposed to obtain rich foreground cues by adaptively fusing low-level spatial details and high-level semantic concepts with the assistance of attention mechanism. After that, the foreground cues are served as the input for the subsequent decoding stage. Considering that there is a lack of two-modal datasets of polyp segmentation, we construct a new dataset with 1,200 images. Extensive experiments on our dataset and three public datasets demonstrate that our CADNet has considerable generalization ability across two-modal data and cross-center data, and obtains comparable results compared with the state-of-the-art methods.
5 June 2025 in International Journal of Pattern Recognition and Artificial Intelligence