Semantic image segmentation
Semantic segmentation is useful in real world applications due to its deep understanding of scene imagery, yet it is one of the most challenging problems in computer vision. The pixel-level nature of the task makes training of models expensive and time-consuming. A popular approach to overcome th...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2023
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/171906 |
| _version_ | 1811688233379037184 |
|---|---|
| author | Melia, Selvira Junita |
| author2 | Lu Shijian |
| author_facet | Lu Shijian Melia, Selvira Junita |
| author_sort | Melia, Selvira Junita |
| collection | NTU |
| description | Semantic segmentation is useful in real world applications due to its deep understanding of scene
imagery, yet it is one of the most challenging problems in computer vision. The pixel-level
nature of the task makes training of models expensive and time-consuming. A popular approach
to overcome this issue is to use a model that was pre-trained on a different dataset. However,
differences in training and test domains cause a reduction in model performance. This project
addresses the problem by focusing on the task of unsupervised domain adaptation in semantic
segmentation. Research into the various novel techniques within the field identified the presence
of hyperparameters specific to techniques being implemented. Adjustments to such
hyperparameters have the potential to provide interesting data on model performance. To
investigate this research gap, this project focuses on the effects of a novel method with a
category contrast technique on a selected baseline method. Experiments were conducted to
measure improvements in model performance based on varying values of the weight of
contrastive loss. It was concluded that a higher weight generally leads to improved model
performance, although variations in scores among different classes suggest future research
directions that focus on fine-tuning technique-specific hyperparameters within the unsupervised
domain adaptation in semantic segmentation field. |
| first_indexed | 2024-10-01T05:28:57Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/171906 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T05:28:57Z |
| publishDate | 2023 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1719062023-11-17T15:37:37Z Semantic image segmentation Melia, Selvira Junita Lu Shijian School of Computer Science and Engineering Shijian.Lu@ntu.edu.sg Engineering::Computer science and engineering Semantic segmentation is useful in real world applications due to its deep understanding of scene imagery, yet it is one of the most challenging problems in computer vision. The pixel-level nature of the task makes training of models expensive and time-consuming. A popular approach to overcome this issue is to use a model that was pre-trained on a different dataset. However, differences in training and test domains cause a reduction in model performance. This project addresses the problem by focusing on the task of unsupervised domain adaptation in semantic segmentation. Research into the various novel techniques within the field identified the presence of hyperparameters specific to techniques being implemented. Adjustments to such hyperparameters have the potential to provide interesting data on model performance. To investigate this research gap, this project focuses on the effects of a novel method with a category contrast technique on a selected baseline method. Experiments were conducted to measure improvements in model performance based on varying values of the weight of contrastive loss. It was concluded that a higher weight generally leads to improved model performance, although variations in scores among different classes suggest future research directions that focus on fine-tuning technique-specific hyperparameters within the unsupervised domain adaptation in semantic segmentation field. Bachelor of Engineering (Computer Science) 2023-11-15T08:25:06Z 2023-11-15T08:25:06Z 2023 Final Year Project (FYP) Melia, S. J. (2023). Semantic image segmentation. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/171906 https://hdl.handle.net/10356/171906 en application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Computer science and engineering Melia, Selvira Junita Semantic image segmentation |
| title | Semantic image segmentation |
| title_full | Semantic image segmentation |
| title_fullStr | Semantic image segmentation |
| title_full_unstemmed | Semantic image segmentation |
| title_short | Semantic image segmentation |
| title_sort | semantic image segmentation |
| topic | Engineering::Computer science and engineering |
| url | https://hdl.handle.net/10356/171906 |
| work_keys_str_mv | AT meliaselvirajunita semanticimagesegmentation |