26 September 2023 in Expert Systems with Applications
6 February 2025 in Expert Systems with Applications
14 March 2025 in IEEE Transactions on Instrumentation and Measurement
Colon polyp segmentation can assist physicians in screening colonoscopy images, which is crucial in preventing colorectal cancer. Due to the limited area occupied by small object polyp objects within images, there is a high risk of overlooking them. That makes it one of the more challenging aspects to address. Additionally, the current segmentation of small object polyps also faces difficulties such as boundary blurring, diverse lesion shapes, and uneven image brightness. While deep learning methods based on convolutional neural networks(CNNs) have been successfully applied to polyp segmentation tasks, three significant challenges persist: (1) limited ability to extract boundary information, (2) inadequate robustness capture of global context information, and (3) insufficient ability for integrating global and local information. To address the issues mentioned above, we propose the Boundary-Guided Global-Local feature fusion network for small polyp segmentation(BGGL-Net), with the following contributions: (1) The local information encoder and boundary feature extraction module leverage convolutional blocks and Laplacian operators to mine feature boundaries and fine-grained detail features. (2) We design a global information fusion module to enhance the model’s representational capacity and acquire rich and accurate global information. (3) Boundary-guided module using a cross-attention mechanism, capturing the inherent relationship between the feature and establishes long-range dependencies between global and low-level features. We enhance boundary accuracy by employing local boundary features to guide the global features, facilitating the effective fusion of global and local information. In the experiment, we compared ten state-of-the-art(SOTA) networks with BGGL-Net. The BGGL-Net achieves the highest segmentation accuracy on small object polyp datasets. Concerning generalization performance, the BGGL-Net outperforms CaraNet by up to 12.4% in the mDice metric on the ETIS-LaribPolypDB* dataset. Our code is available at https://github.com/shenjoyao/BGGL-Net.
21 August 2025 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Road segmentation from remote sensing images is crucial in various applications, including map navigation, urban planning, vegetation change analysis and disaster assessment. However, problems such as minor inter-class differences, shadow masking, and narrow or interrupted roads frequently occur, making it difficult for most methods relying on convolution or self-attention to quickly and comprehensively extract these road objects. In this paper, we propose a Residual Complex Fourier Network (RCFNet), which utilizes Fourier Encoding Operators (FEO) to quickly and accurately extract roads. Firstly, a Fourier Encoder Module (FEM) is proposed to extract road information using complex-valued features in the frequency domain. Secondly, a Road Information Mix (RIM) module is introduced to align frequency-domain features across different scales. Finally, a multi-feature fusion strategy is proposed to calibrate multi-level features from the decoder to enhance road segmentation results. Extensive experiments on two public datasets, the DeepGlobe road dataset and the Massachusetts road dataset, have demon-strated that RCFNet achieves notable performance with minimal computational cost (FLOPs).