Learning with limited anotated data for medical image diagnosis
Project Description

This project will cover semi-supervised learning, self-supervised/unsupervised learning, few-shot learning for medical image diagnosis. The main goal is to make the AI algorithm more intelligent by learning with unlabeled data or learning with few labeled data.

Supervisor
LI Xiaomeng
Quota
10
Course type
UROP1000
UROP1100
UROP2100
UROP3100
UROP4100
Applicant's Roles

1, Students will read papers, conduct experiments, and discuss ideas with supervisors.
2, Students should at least have a strong interest in research in AI and computer vision.
3, A weekly meeting is required.

Applicant's Learning Objectives

1. Students are expected to have a deep understanding of semi-supervised, unsupervised and few-shot learning.
2. They are expected to know the existing research, including both the advantages and disadvantages.
3, They should be able to have their own ideas for solving cutting-edge problems and have research outcomes.

Complexity of the project
Moderate