Focus U-Net: A novel dual attention-gated CNN for polyp segmentation during colonoscopy

• Published in: Computers in biology and medicine 2021-10, Vol.137, p.104815-104815, Article 104815
• Main Authors: Yeung, Michael, Sala, Evis, Schönlieb, Carola-Bibiane, Rundo, Leonardo
• Format: Article
• Language: English
• Subjects: Ablation; Accuracy; Artificial intelligence; Artificial neural networks; Attention mechanisms; Cancer; Cancer screening; Colon; Colonoscopy; Colorectal cancer; Colorectal carcinoma; Computer-aided diagnosis; Datasets; Deep learning; Focus U-Net; Image processing; Image segmentation; Localization; Loss function; Machine learning; Medical diagnosis; Medical imaging; Medical prognosis; Neural networks; Performance evaluation; Polyp segmentation; Polyps; Screening; Semantics; Tumors
• ISSN: 0010-4825; 1879-0534
• DOI: 10.1016/j.compbiomed.2021.104815

Colonoscopy remains the gold-standard screening for colorectal cancer. However, significant miss rates for polyps have been reported, particularly when there are multiple small adenomas. This presents an opportunity to leverage computer-aided systems to support clinicians and reduce the number of polyps missed. In this work we introduce the Focus U-Net, a novel dual attention-gated deep neural network, which combines efficient spatial and channel-based attention into a single Focus Gate module to encourage selective learning of polyp features. The Focus U-Net incorporates several further architectural modifications, including the addition of short-range skip connections and deep supervision. Furthermore, we introduce the Hybrid Focal loss, a new compound loss function based on the Focal loss and Focal Tversky loss, designed to handle class-imbalanced image segmentation. For our experiments, we selected five public datasets containing images of polyps obtained during optical colonoscopy: CVC-ClinicDB, Kvasir-SEG, CVC-ColonDB, ETIS-Larib PolypDB and EndoScene test set. We first perform a series of ablation studies and then evaluate the Focus U-Net on the CVC-ClinicDB and Kvasir-SEG datasets separately, and on a combined dataset of all five public datasets. To evaluate model performance, we use the Dice similarity coefficient (DSC) and Intersection over Union (IoU) metrics. Our model achieves state-of-the-art results for both CVC-ClinicDB and Kvasir-SEG, with a mean DSC of 0.941 and 0.910, respectively. When evaluated on a combination of five public polyp datasets, our model similarly achieves state-of-the-art results with a mean DSC of 0.878 and mean IoU of 0.809, a 14% and 15% improvement over the previous state-of-the-art results of 0.768 and 0.702, respectively. This study shows the potential for deep learning to provide fast and accurate polyp segmentation results for use during colonoscopy. The Focus U-Net may be adapted for future use in newer non-invasive colorectal cancer screening and more broadly to other biomedical image segmentation tasks similarly involving class imbalance and requiring efficiency. [Display omitted] •Automatic polyp segmentation can support clinicians to reduce polyp miss rates.•Focus U-Net combines efficient spatial and channel attention into a Focus Gate.•Focus Gate uses tunable focal parameter to control degree of background suppression.•Deep supervision and new Hybrid Focal loss function further improve performance.•Focus U-