Generalizable Polyp Segmentation via Randomized Global Illumination Augmentation

Accuratelysegmenting polyps from colonoscopy images is essential for diagnosing colorectal cancer. Despite the tremendous success of the deep convolutional neural networks in automatic polyp segmentation, it suffers from domain shift issues, where the trained model yields performance deterioration on unseen test datasets. This article proposes an illumination enhancement-based domain generalization approach to improve the generalization capability of the model on unseen test datasets and alleviate this issue. In particular, an image decomposition module (IDM) was developed to separate colonoscopy images into reflectance, local, and global illumination components. An illumination transform module (ITM) was proposed to augment images with different global illuminations by synthesizing target-like global illumination maps. A novel illumination variance insensitiveness (IViSen) is also introduced to evaluate the robustness of the model against illumination disturbance. IViSen is easy to compute and correlates well with model generalizability. The segmentation performance of the proposed model on four colonoscopy datasets was examined: CVC-ClinicDB, CVC-ColonDB, ETIS-Larib, and Kvasir-SEG. The method outperformed the competitive methods when tested on unseen domains. In particular, the proposed approach yielded 60. 82% and 53. 19% in terms of mean Dice and IoU, respectively, with 2. 06% and 2. 31% improvements.

Authors

• Zuyu Zhang
• Yan Li
• Byeong-Seok Shin

Keywords

• Lighting
• Image segmentation
• Colonoscopy
• Reflectivity
• Task analysis
• Robustness
• Medical diagnostic imaging
• Polyp Segmentation
• Colorectal Cancer
• Deep Neural Network
• Domain Shift
• Global Map
• Generalization Capability
• Segmentation Performance
• Local Components
• Domain Generalization
• Image Augmentation
• Illumination Variations
• Image Decomposition
• Global Component
• Reflection Component
• Mean Intersection Over Union
• Generalization Capability Of The Model
• Unseen Domains
• Medical Imaging
• Structural Information
• Data Augmentation
• Source Domain
• Jensen-Shannon Divergence
• Target Domain
• Illumination Changes
• Baseline Methods
• Medical Image Segmentation
• Image Pairs
• Generalization Performance Of The Model
• Illumination Conditions
• Original Input Image