MIT & MITM 12/18cp Projects
Perception-based objective evaluation of image segmentation
Summary:
This project will provide a holistic investigation of objective criteria for evaluation of image segmentation results,
aiming to find out those which are able to produce the results consistent with human perception.
Synopsis:
Image segmentation usually refers to subdividing an image into its constituent parts and extracting regions of interest.
The plethora of image segmentation techniques in the literature and their wide use in a variety of applications makes
increasingly important the need for the objective evaluation of segmentation results. Besides supporting the selection
of suitable segmentation tools, segmentation evaluation can also serve in joint segmentation/evaluation schemes so as
to iteratively improve the segmentation performance. Although a large number of evaluation criteria have been proposed,
many of them cannot produce the assessment which agrees with human perception.
This project aims to make a comprehensive and systematic investigation of segmentation evaluation techniques. We will find
out those objective evaluation criteria which are consistent with human perception through analysing the relevance between
objective and subjective evaluation. The output of this project will facilitate image segmentation and analysis.
By finishing this project, students will learn some technologies in digital image processing and multimedia and be more
experienced in programming and problem solving.
Keywords: Image analyse, Image segmentation, Segmentation evaluation
Supervisors: Prof. David Dagan Feng, Dr. Yong Xia
Research Locations: Level 3, School of IT building J12
