Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

CENet: Context Enhancement Network for Medical Image Segmentation

Afshin Bozorgpour, Sina Ghorbani Kolahi, Reza Azad, Ilker Hacihaliloglu, Dorit Merhof

TL;DR

Medical image segmentation across diverse domains requires preserving boundary details while integrating local and global context. CENet addresses this with a Pyramid Vision Transformer V2 backbone and two novel modules: Dual Selective Enhancement Block (DSEB) to enrich skip connections and Context Feature Attention Module (CFAM) to refine multiscale features in the decoder. The approach yields state-of-the-art performance on radiology and dermoscopy datasets, improving boundary preservation and multi-organ segmentation robustness. The work provides a practical, interpretable framework with publicly available code that can advance clinical image analysis tasks across modalities.

Abstract

Medical image segmentation, particularly in multi-domain scenarios, requires precise preservation of anatomical structures across diverse representations. While deep learning has advanced this field, existing models often struggle with accurate boundary representation, variability in organ morphology, and information loss during downsampling, limiting their accuracy and robustness. To address these challenges, we propose the Context Enhancement Network (CENet), a novel segmentation framework featuring two key innovations. First, the Dual Selective Enhancement Block (DSEB) integrated into skip connections enhances boundary details and improves the detection of smaller organs in a context-aware manner. Second, the Context Feature Attention Module (CFAM) in the decoder employs a multi-scale design to maintain spatial integrity, reduce feature redundancy, and mitigate overly enhanced representations. Extensive evaluations on both radiology and dermoscopic datasets demonstrate that CENet outperforms state-of-the-art (SOTA) methods in multi-organ segmentation and boundary detail preservation, offering a robust and accurate solution for complex medical image analysis tasks. The code is publicly available at https://github.com/xmindflow/cenet.

CENet: Context Enhancement Network for Medical Image Segmentation

TL;DR

Medical image segmentation across diverse domains requires preserving boundary details while integrating local and global context. CENet addresses this with a Pyramid Vision Transformer V2 backbone and two novel modules: Dual Selective Enhancement Block (DSEB) to enrich skip connections and Context Feature Attention Module (CFAM) to refine multiscale features in the decoder. The approach yields state-of-the-art performance on radiology and dermoscopy datasets, improving boundary preservation and multi-organ segmentation robustness. The work provides a practical, interpretable framework with publicly available code that can advance clinical image analysis tasks across modalities.

Abstract

Medical image segmentation, particularly in multi-domain scenarios, requires precise preservation of anatomical structures across diverse representations. While deep learning has advanced this field, existing models often struggle with accurate boundary representation, variability in organ morphology, and information loss during downsampling, limiting their accuracy and robustness. To address these challenges, we propose the Context Enhancement Network (CENet), a novel segmentation framework featuring two key innovations. First, the Dual Selective Enhancement Block (DSEB) integrated into skip connections enhances boundary details and improves the detection of smaller organs in a context-aware manner. Second, the Context Feature Attention Module (CFAM) in the decoder employs a multi-scale design to maintain spatial integrity, reduce feature redundancy, and mitigate overly enhanced representations. Extensive evaluations on both radiology and dermoscopic datasets demonstrate that CENet outperforms state-of-the-art (SOTA) methods in multi-organ segmentation and boundary detail preservation, offering a robust and accurate solution for complex medical image analysis tasks. The code is publicly available at https://github.com/xmindflow/cenet.

Paper Structure

This paper contains 8 sections, 5 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Method
  3. Dual Selective Enhancement Block (DSEB)
  4. Contextual Feature Attention Module (CFAM)
  5. Experiments and Results
  6. Datasets and Implementation Details
  7. Results
  8. Conclusion

Figures (4)

  • Figure 1: The Contextual Enhancement Network (CENet) uses a pretrained encoder to generate multi-resolution features, processed by the DSEB as skip connections. The decoder refines these features via the CFAM, which includes a CCU, MCA, wNLB, and MLP.
  • Figure 2: Visual comparison of the proposed method versus others on the Synapse dataset.
  • Figure 3: Qualitative comparison of CENet and previous methods across skin benchmarks.
  • Figure 4: Feature visualization in CENet: first row shows an ACDC sample, second row a PH$^2$ example