MarineInst: foundation model for marine image analysis with instance visual description
Project Description
MarineInst is a foundation model for the analysis of the marine realms with instance visual description, which outputs instance masks and captions for marine object instances. To generate informative and detailed semantic instance captions, we use vision-language models to produce semantic richness with various granularities.
Supervisor
YEUNG, Sai Kit
Quota
5
Course type
UROP1000
UROP1100
UROP2100
UROP3100
UROP3200
UROP4100
Applicant's Roles
Help collect the dataset, which contains a wide spectrum of marine images with high-quality semantic instance masks.
Applicant's Learning Objectives
Students will learn the open-vocabulary instance segmentation algorithms. Learn the basics of vision-language models.
Complexity of the project
Moderate