Edge-Aware Mirror Network for Camouflaged Object Detection

Dongyue Sun,Shiyao Jiang,Lin Qi

2023-07-08

Abstract:Existing edge-aware camouflaged object detection (COD) methods normally output the edge prediction in the early stage. However, edges are important and fundamental factors in the following segmentation task. Due to the high visual similarity between camouflaged targets and the surroundings, edge prior predicted in early stage usually introduces erroneous foreground-background and contaminates features for segmentation. To tackle this problem, we propose a novel Edge-aware Mirror Network (EAMNet), which models edge detection and camouflaged object segmentation as a cross refinement process. More specifically, EAMNet has a two-branch architecture, where a segmentation-induced edge aggregation module and an edge-induced integrity aggregation module are designed to cross-guide the segmentation branch and edge detection branch. A guided-residual channel attention module which leverages the residual connection and gated convolution finally better extracts structural details from low-level features. Quantitative and qualitative experiment results show that EAMNet outperforms existing cutting-edge baselines on three widely used COD datasets. Codes are available at <a class="link-external link-https" href="https://github.com/sdy1999/EAMNet" rel="external noopener nofollow">this https URL</a>.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

### Problems Addressed by the Paper

This paper proposes a new framework—Edge-aware Mirror Network (EAMNet), aimed at addressing the challenging issues in the task of Camouflaged Object Detection (COD).

#### Core Issues
1. **Edge priors introducing incorrect foreground-background segmentation**:
- Existing COD methods typically output edge predictions at an early stage. However, due to the high visual similarity between camouflaged objects and the background, early predicted edge priors often lead to incorrect foreground-background segmentation, contaminating features in subsequent segmentation tasks.

2. **Insufficient utilization of low-frequency texture information**:
- Current methods pay insufficient attention to low-level features that retain low-frequency texture information, resulting in inaccurate edge predictions.

3. **Lack of accuracy consideration when injecting edge priors into segmentation features**:
- There is an overemphasis on injecting edge priors into segmentation features, but the accuracy of the early-generated edge predictions themselves is often overlooked, which can lead to incorrect foreground predictions in the segmentation results.

#### Solutions
EAMNet addresses the above issues through the following approaches:

1. **Two-branch architecture**:
- EAMNet includes an edge detection branch and a segmentation branch, where the two branches guide each other to jointly optimize the segmentation results.

2. **Cross-refinement process**:
- The Edge-induced Integrity Aggregation (EIA) module and the Segmentation-induced Edge Aggregation (SEA) module are proposed to achieve cross-guidance between edge detection and object segmentation.

3. **Low-level feature extraction for structural details**:
- The Guided-residual Channel Attention (GCA) module is introduced to utilize low-level features to extract structural details of camouflaged objects.

Through these designs, EAMNet achieves significantly better performance than existing methods on three widely used COD datasets.