Abstract:Background Colorectal cancer is one of the commonest malignancies which threaten the health of human. Most of the colorectal cancers arise from adenomatous polyps, screening for this cancer is effective not only for early detection but also for prevention. Endoscopic examination is an important way to reveal polyps. At present, the main limitation of colonoscopy is the high polyp miss rate. This study aims to improve the accuracy of detection and save doctors’ time with deep learning techniques. Methods The data we used is a combination of CVC-ColonDB and CVC-ClinicDB which contains 912 images with associated polyps. We treat the rest part of a colonoscopy image except polys as background. Convolutional networks are applied to polyp localization and segmentation. There are three different architectures which include FCN, SegNet and modified ResNet. We trained convolutional networks with these architectures respectively and evaluated these models based on Intersection over Union (IoU) and accuracy. Results The segmentation results on the test set are reported in table 1 and the accuracy of three architectures is greater than 90%. Results achieved in SegNet and modified ResNet are very encouraging, with a IoU higher than 50% for polyps. IDDF2018-ABS-0260 Figure 1 shows qualitative results of three architectures. It can be seen that all of them recognise the location of polyp successfully and rows 2 and 3 show finer boundaries. The results are reported on test setAbstract IDDF2018-ABS-0260 Table 1 Endoluminal scene semantic segmentation results for different architectures Architecture IoU background IoU polyp IoU mean Acc mean FCN8 94.024 46.99 70.51 94.32 SegNet 94.897 53.795 74.346 95.183 ResNet with changes 96.029 64.0307 80.0299 96.291 Abstract IDDF2018-ABS-0260 Figure 1 Examples of predictions: (a) FCN8 model; (b) SegNet model; (c) modified ResNet model. From left to right, each column represents ground truth, prediction and a translucent prediction image Conclusions Compare to traditional time-consuming hand-crafted segmentation methods, when considering polyp segmentation, approaches based on deep learning are time-saving and effective, showing good results in colonoscopy images. Given that three architectures we mentioned above not only performs well but also allows for nearly real-time processing, it has a great potential in polyp localization and segmentation.

FCN-Transformer Feature Fusion for Polyp Segmentation

Fusing Transformer and FCN for Polyp Segmentation

Polyp Segmentation With the FCB-SwinV2 Transformer

FCB-SwinV2 Transformer for Polyp Segmentation

NA-segformer: A multi-level transformer model based on neighborhood attention for colonoscopic polyp segmentation

Improving Polyp Segmentation with Boundary-Assisted Guidance and Cross-Scale Interaction Fusion Transformer Network

PolySegNet: improving polyp segmentation through swin transformer and vision transformer fusion

Cooperation Learning Enhanced Colonic Polyp Segmentation Based on Transformer-CNN Fusion

UMF-Net: A UNet-based Multi-Branch Feature Fusion Network for Colon Polyp Segmentation

A lighter hybrid feature fusion framework for polyp segmentation

DBMF: Dual Branch Multiscale Feature Fusion Network for polyp segmentation

Multi‐scale nested UNet with transformer for colorectal polyp segmentation

SegT: A Novel Separated Edge-guidance Transformer Network for Polyp Segmentation

TranSEFusionNet: Deep fusion network for colorectal polyp segmentation

W-MAFormer: W-shaped Multi-Attention Assisted Transformer for Polyp Segmentation

Transformer-Enhanced Iterative Feedback Mechanism for Polyp Segmentation

IDDF2018-ABS-0260 Deep Learning for Polyp Segmentation

SAEFormer: stepwise attention emphasis transformer for polyp segmentation

Utilizing adaptive deformable convolution and position embedding for colon polyp segmentation with a visual transformer

A Segmentation Algorithm of Colonoscopy Images Based on Multi-Scale Feature Fusion

DEMF-Net: A dual encoder multi-scale feature fusion network for polyp segmentation