Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ODC-SA Net: Orthogonal Direction Enhancement and Scale Aware Network for Polyp Segmentation

Chenhao Xu, Yudian Zhang, Kaiye Xu, Haijiang Zhu

TL;DR

An Orthogonal Direction Enhancement and Scale Aware Network (ODC-SA Net) for polyp segmentation and the Multi-scale Fusion Attention mechanism is proposed to emphasize scale changes in both spatial and channel dimensions, enhancing the segmentation accuracy for polyps of varying sizes.

Abstract

Accurate polyp segmentation is crucial for the early detection and prevention of colorectal cancer. However, the existing polyp detection methods sometimes ignore multi-directional features and drastic changes in scale. To address these challenges, we design an Orthogonal Direction Enhancement and Scale Aware Network (ODC-SA Net) for polyp segmentation. The Orthogonal Direction Convolutional (ODC) block can extract multi-directional features using transposed rectangular convolution kernels through forming an orthogonal feature vector basis, which solves the issue of random feature direction changes and reduces computational load. Additionally, the Multi-scale Fusion Attention (MSFA) mechanism is proposed to emphasize scale changes in both spatial and channel dimensions, enhancing the segmentation accuracy for polyps of varying sizes. Extraction with Re-attention Module (ERA) is used to re-combinane effective features, and Structures of Shallow Reverse Attention Mechanism (SRA) is used to enhance polyp edge with low level information. A large number of experiments conducted on public datasets have demonstrated that the performance of this model is superior to state-of-the-art methods.

ODC-SA Net: Orthogonal Direction Enhancement and Scale Aware Network for Polyp Segmentation

TL;DR

An Orthogonal Direction Enhancement and Scale Aware Network (ODC-SA Net) for polyp segmentation and the Multi-scale Fusion Attention mechanism is proposed to emphasize scale changes in both spatial and channel dimensions, enhancing the segmentation accuracy for polyps of varying sizes.

Abstract

Accurate polyp segmentation is crucial for the early detection and prevention of colorectal cancer. However, the existing polyp detection methods sometimes ignore multi-directional features and drastic changes in scale. To address these challenges, we design an Orthogonal Direction Enhancement and Scale Aware Network (ODC-SA Net) for polyp segmentation. The Orthogonal Direction Convolutional (ODC) block can extract multi-directional features using transposed rectangular convolution kernels through forming an orthogonal feature vector basis, which solves the issue of random feature direction changes and reduces computational load. Additionally, the Multi-scale Fusion Attention (MSFA) mechanism is proposed to emphasize scale changes in both spatial and channel dimensions, enhancing the segmentation accuracy for polyps of varying sizes. Extraction with Re-attention Module (ERA) is used to re-combinane effective features, and Structures of Shallow Reverse Attention Mechanism (SRA) is used to enhance polyp edge with low level information. A large number of experiments conducted on public datasets have demonstrated that the performance of this model is superior to state-of-the-art methods.
Paper Structure (21 sections, 7 equations, 7 figures, 2 tables)

This paper contains 21 sections, 7 equations, 7 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Orthogonal Direction Convolutional (ODC) Block
  4. Principle.
  5. Operation.
  6. Multi-scale Fusion Attention (MSFA)
  7. Spatial Scalable Enhancement Mechanism (S2E).
  8. Channel Scalable Attention Mechanism (CSA).
  9. Residual Fusion Attention (RFA).
  10. Extraction with Re-attention Module (ERA)
  11. Shallow Reverse Attention Mechanism (SRA)
  12. Loss Function
  13. Experiment
  14. Details
  15. Dataset.
  16. ...and 6 more sections

Figures (7)

  • Figure 1: Structure of Orthogonal Direction Enhancement and Scale Aware Network.
  • Figure 2: Structure of the key sectors of Orthogonal Direction Convolutional (ODC) block, which can be also described as Eq.(\ref{['eq3']})-(\ref{['eq5']}).
  • Figure 3: Structures of Multi-scale Fusion Attention (MSFA) and Extraction with Re-attention Module (ERA).
  • Figure 4: Structures of Shallow Reverse Attention Mechanism (SRA).
  • Figure 5: Qualitative results of different methods.
  • ...and 2 more figures