Transformer-Based Disease Identification for Small-Scale Imbalanced Capsule Endoscopy Dataset

Transformer-Based Disease Identification for Small-Scale Imbalanced Capsule Endoscopy Dataset

Vision Transformer (ViT) is emerging as a new leader in computer vision with its outstanding performance in many tasks (e.g., ImageNet-22k, JFT-300M). However, the success of ViT relies on pretraining on large datasets. It is difficult for us to use ViT to train from scratch on a small-scale imbalan...

Full description

Bibliographic Details
Main Authors: Long Bai, Liangyu Wang, Tong Chen, Yuanhao Zhao, Hongliang Ren
Format: Article
Language:English
Published: MDPI AG 2022-08-01
Series:Electronics
Subjects:
vision transformer
wireless capsule endoscopy
disease identification
small-scale and imbalance dataset
pooling
Online Access:https://www.mdpi.com/2079-9292/11/17/2747

Internet

https://www.mdpi.com/2079-9292/11/17/2747

Similar Items