SR-AttNet: An Interpretable Stretch–Relax Attention based Deep Neural Network for Polyp Segmentation in Colonoscopy Images

• Published in: Computers in biology and medicine Vol. 160; p. 106945
• Main Authors: Alam, Md. Jahin, Fattah, Shaikh Anowarul
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.06.2023; Elsevier Limited
• Subjects: Artificial neural networks; Attention; CNN; Coders; Cognitive ability; Colon; Colonic Polyps - diagnostic imaging; Colonoscopy; Deep learning; Filtration; Humans; Image processing; Image Processing, Computer-Assisted; Image segmentation; Inspection; Interpretable; Machine learning; Neural networks; Neural Networks, Computer; Optimization; Polyp segmentation; Polyps; Deep learning; CNN; Colonoscopy; Polyp segmentation; Interpretable; Attention
• ISSN: 0010-4825; 1879-0534
• DOI: 10.1016/j.compbiomed.2023.106945

Colorectal polyp is a common structural gastrointestinal (GI) anomaly, which can in certain cases turn malignant. Colonoscopic image inspection is, thereby, an important step for isolating the polyps as well as removing them if necessary. However, the process is around 30-60 min long and inspecting each image for polyps can prove to be a tedious task. Hence, an automatic computerized process for efficient and accurate polyp isolation can be a useful tool. In this study, a deep learning network is introduced for colorectal polyp segmentation. The network is based on an encoder–decoder architecture, however, having both un-dilated and dilated filtering in order to extract both near and far local information as well as perceive image depth. Four-fold skip-connections exist between each spatial encoder–decoder due to both type of filtering and a ‘Feature-to-Mask’ pipeline processes the decoded dilated and un-dilated features for final prediction. The proposed network implements a ‘Stretch–Relax’ based attention system, SR-Attention, to generate high variance spatial features in order to obtain useful attention masks for cognitive feature selection. From this ‘Stretch–Relax’ attention based operation, the network is termed as ‘SR-AttNet’. Training and optimization is performed on four different datasets, and inference has been done on five (Kvasir-SEG, CVC-ClinicDB, CVC-Colon, ETIS-Larib, EndoCV2020); all of which output higher Dice-score compared to state-of-the-art and existing networks. The efficacy and interpretability of SR-Attention is also demonstrated based on quantitative variance. In consequence, the proposed SR-AttNet can be considered for an automated and general approach for polyp segmentation during colonoscopy. •Utilization of both un-dilated and dilated filters in the encoder and decoder.•Stretch–Relax type Attention system within encoder and decoder pipeline.•Additional Feature-to-Mask Pipeline for feature aggregation and prediction.•Generalizability testing of the proposed SR-AttNet.•Interpretation of the Attention system, visually and quantitatively.