Automatic detection of erosions and ulcerations in wireless capsule endoscopy images based on a deep convolutional neural network

Clinical usefulness of a deep learning‐based system as the first screening on small‐bowel capsule endoscopy reading

“…These images were not included in the training set for developing our CNN system. 11 In that period, capsule endoscopy was carried out using the Pillcam SB3 device (Medtronic, Minneapolis, MN, USA).…”

“…We previously developed an artificial intelligence (AI)based diagnostic system based on a deep neural network architecture called the Single Shot MultiBox Detector (SSD, https://arxiv.org/abs/1512.02325) without altering its algorithm. 11 SSD is a deep CNN that consists of 16 or more layers. The CNN system was reported to automatically detect erosions or ulcerations with high accuracy.…”

“…We previously developed a convolutional neural network (CNN) system using a deep learning method, which could automatically detect mucosal breaks with high accuracy. 11 However, the usefulness of the proposed CNN system in clinical practice is still unknown. In addition, we wonder whether the usefulness may be different according to the reading level of endoscopists.…”

Artificial intelligence using a convolutional neural network for automatic detection of small‐bowel angioectasia in capsule endoscopy images

“…A CNN, a type of deep-learning method with multilayer perceptrons designed to use minimal preprocessing, was recently reported as being highly beneficial in the field of endoscopy. Aoki et al 32 reported that a CNN-based diagnostic program was useful for detecting erosion/ulcer of CE still images and was helpful for physicians. They reported that the sensitivity, specificity, and accuracy of CNN were 88.2%, 90.9%, and 90.8%, respectively, at a cut-off value of 0.481 for the probability score.…”

Evaluation of current status and near future perspectives of capsule endoscopy: Summary of Japan Digestive Disease Week 2019

“…T HE APPLICATION OF artificial intelligence (AI) using deep learning-based CE has been reported for automatic detection of angioectasia, erosions, ulcerations, and tumors. 7,8 A computer-aided diagnosis algorithm would be a valuable asset with regard to the time-consuming task of reviewing multiple CE images. Such tools will aid physicians in facilitating and accelerating the CE reviewing process.…”