Bridge Cheating Detection based on Computer Vision and Machine Learning

Primary supervisor

Qiuhong Ke

This project aims to develop a method based on computer vision and machine learning techniques for bridge cheating detection. Bridge is a trick-taking card game using a standard 52-card deck. In its basic format, it is played by four players in two competing partnerships, with partners sitting opposite each other around a table. Cheating in bridge refers to a deliberate violation of the rules of the game of bridge or other unethical behaviour that is intended to give an unfair advantage to a player or team. Cheating can occur in many forms such as conveying information to a partner by means of a pre-arranged illegal signal, viewing the opponents' cards in a board prior to their arrival at the table and altering the records as to the results of a board. This project aims to identify if a pair of partners convey information using card orientation, hand gestures or other audio information such as unnatural noise. To this end, this project will need to recognise fine-grained visual and audio information to further investigate if there are anomalous patterns during the game, which can belong to cheating.

Student cohort

Single Semester
Double Semester

Aim/outline

1) Recognise card orientation

2) Recognise other card information

3) Identify anomalous patterns based on multi-modal information, including card information, hand gesture and audio.

Required knowledge

This project will involve video processing and analysis based on computer vision and machine learning. Ideally, the students should have computer vision and machine learning knowledge, as well as good programming skills. Relevant experience on video processing will be preferred.