Dynamic texture classification

• Main Author: Chockalingam, Muthiah
• Other Authors: Deepu Rajan
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision; Engineering::Computer science and engineering::Computing methodologies::Pattern recognition
• Online Access: https://hdl.handle.net/10356/156532
• Institution: Nanyang Technological University

Video Object Segmentation is a segment under the field of computer vision, and it recently has been garnering a lot of attention as to its applicability in solving many real-life problems. One such problem is the effective navigation of ships, by using video objects segmentation to effectively segment the target object (i.e., water) to identify clear regions of water for ships to sail through and avoid any possible obstacles on their course. The biggest problem with creating such a video object segmentation model would be the fact that water has a very tricky appearance. Its appearance and texture are highly dynamic as they change very quickly, sometimes even between frames itself, due to factors such as illumination, ripples, and waves. Therefore, this student’s aim is to analyse the existing video object segmentation methods, identify the most suitable one for dynamic texture classification (i.e., tracking object with a dynamic appearance), and test it with a dataset that is representative of the waters that ships sail through. Results are then recorded to evaluate the effectiveness of that particular model with regards to the objective of aiding a ship’s navigation.