2MGAS-Net: multi-level multi-scale gated attentional squeezed network for polyp segmentation

Signal Image and Video Processing, Journal Year: 2024, Volume and Issue: 18(6-7), P. 5377 - 5386

Published: May 10, 2024

Language: Английский

---

Lightweight medical image segmentation network with multi-scale feature-guided fusion

Computers in Biology and Medicine, Journal Year: 2024, Volume and Issue: 182, P. 109204 - 109204

Published: Oct. 3, 2024

Language: Английский

---

A survey on deep learning for polyp segmentation: techniques, challenges and future trends

Visual Intelligence, Journal Year: 2025, Volume and Issue: 3(1)

Published: Jan. 3, 2025

Language: Английский

---

Multi-Bottleneck progressive propulsion network for medical image semantic segmentation with integrated macro-micro dual-stage feature enhancement and refinement

Expert Systems with Applications, Journal Year: 2024, Volume and Issue: 252, P. 124179 - 124179

Published: May 14, 2024

Language: Английский

---

SCSONet: spatial-channel synergistic optimization net for skin lesion segmentation

Frontiers in Physics, Journal Year: 2024, Volume and Issue: 12

Published: March 20, 2024

In the field of computer-assisted medical diagnosis, developing image segmentation models that are both accurate and capable real-time operation under limited computational resources is crucial. Particularly for skin disease segmentation, construction such lightweight must balance cost efficiency, especially in environments with computing power, memory, storage. This study proposes a new network designed specifically aimed at significantly reducing number parameters floating-point operations while ensuring performance. The proposed ConvStem module, full-dimensional attention, learns complementary attention weights across all four dimensions convolution kernel, effectively enhancing recognition irregularly shaped lesion areas, model’s parameter count burden, thus promoting model lightweighting performance improvement. SCF Block reduces feature redundancy through spatial channel fusion, lowering improving results. paper validates effectiveness robustness SCSONet on two public datasets, demonstrating its low resource requirements. https://github.com/Haoyu1Chen/SCSONet .

Language: Английский

---

AFN-Net: Adaptive Fusion Nucleus Segmentation Network Based on Multi-Level U-Net

Sensors, Journal Year: 2025, Volume and Issue: 25(2), P. 300 - 300

Published: Jan. 7, 2025

The task of nucleus segmentation plays an important role in medical image analysis. However, due to the challenge detecting small targets and complex boundaries datasets, traditional methods often fail achieve satisfactory results. Therefore, a novel method based on U-Net architecture is proposed overcome this issue. Firstly, we introduce Weighted Feature Enhancement Unit (WFEU) encoder decoder fusion stage U-Net. By assigning learnable weights different feature maps, network can adaptively enhance key features suppress irrelevant or secondary features, thus maintaining high-precision performance backgrounds. In addition, further improve under resolution designed Double-Stage Channel Optimization Module (DSCOM) first two layers model. This DSCOM effectively preserves high-resolution information improves accuracy boundary regions through multi-level convolution operations channel optimization. Finally, Adaptive Fusion Loss (AFLM) that balances lossy by dynamically adjusting weights, thereby improving model's region consistency while classification accuracy. experimental results 2018 Data Science Bowl demonstrate that, compared state-of-the-art models, our shows significant advantages multiple metrics. Specifically, model achieved IOU score 0.8660 Dice 0.9216, with parameter size only 7.81 M. These illustrate paper not excels shapes but also significantly enhances overall at lower computational costs. research offers new insights references for design future tasks.

Language: Английский

---

BMANet: Boundary-guided multi-level attention network for polyp segmentation in colonoscopy images

Biomedical Signal Processing and Control, Journal Year: 2025, Volume and Issue: 105, P. 107524 - 107524

Published: Jan. 30, 2025

Language: Английский

---

DCATNet: polyp segmentation with deformable convolution and contextual-aware attention network

BMC Medical Imaging, Journal Year: 2025, Volume and Issue: 25(1)

Published: April 14, 2025

Polyp segmentation is crucial in computer-aided diagnosis but remains challenging due to the complexity of medical images and anatomical variations. Current state-of-the-art methods struggle with accurate polyp variability size, shape, texture. These factors make boundary detection challenging, often resulting incomplete or inaccurate segmentation. To address these challenges, we propose DCATNet, a novel deep learning architecture specifically designed for DCATNet U-shaped network that combines ResNetV2-50 as an encoder capturing local features Transformer modeling long-range dependencies. It integrates three key components: Geometry Attention Module (GAM), Contextual Gate (CAG), Multi-scale Feature Extraction (MSFE) block. We evaluated on five public datasets. On Kvasir-SEG CVC-ClinicDB, model achieved mean dice scores 0.9351 0.9444, respectively, outperforming previous (SOTA) methods. Cross-validation further demonstrated its superior generalization capability. Ablation studies confirmed effectiveness each component DCATNet. Integrating GAM, CAG, MSFE effectively improves feature representation fusion, leading precise reliable results. findings underscore DCATNet's potential clinical application can be used wide range image tasks.

Language: Английский

---

MGMFormer: Multi‐Scale Attentional Medical Image Segmentation Network for Semantic Feature Enhancement

International Journal of Imaging Systems and Technology, Journal Year: 2025, Volume and Issue: 35(3)

Published: April 18, 2025

ABSTRACT Multi‐scale feature extraction is important for the accurate segmentation of different lesion areas. In order to solve problem false cut and missing in practical applications due difficulty extracting semantic information from existing technologies, we proposed a multi‐scale attention network framework based on enhancement, MGMFormer. Taking advantage mechanism enhance features, encoder decoder are composed joint learning, arbitrary sampling, global adaptive calibration modules. It makes more focused fine structure, so as effectively deal with reduced accuracy caused by modal heterogeneity. At same time, it solves lack expression ability when deals complex texture information. We evaluated performance MGMFormer eight datasets, BraTS, Sypanse, ACDC, ISIC, Kvasir‐SEG, CAMUS, CHNCXR, Glas, particular, outperformed most algorithms.

Language: Английский

---

ETC-Net: an efficient collaborative transformer and convolutional network combining edge constraints for medical image segmentation

Evolving Systems, Journal Year: 2025, Volume and Issue: 16(2)

Published: May 3, 2025

Language: Английский

---