Bootstrapping the performance of webly supervised semantic segmentation
Fully supervised methods for semantic segmentation require pixel-level class masks to train, the creation of which is expensive in terms of manual labour and time. In this work, we focus on weak supervision, developing a method for training a high-quality pixel-level classifier for semantic segmenta...
| Main Authors: | , , , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/142802 |
| _version_ | 1826127039159074816 |
|---|---|
| author | Shen, Tong Lin, Guosheng Shen, Chunhua Reid, Ian |
| author2 | School of Computer Science and Engineering |
| author_facet | School of Computer Science and Engineering Shen, Tong Lin, Guosheng Shen, Chunhua Reid, Ian |
| author_sort | Shen, Tong |
| collection | NTU |
| description | Fully supervised methods for semantic segmentation require pixel-level class masks to train, the creation of which is expensive in terms of manual labour and time. In this work, we focus on weak supervision, developing a method for training a high-quality pixel-level classifier for semantic segmentation, using only image-level class labels as the provided ground-truth. Our method is formulated as a two-stage approach in which we first aim to create accurate pixel-level masks for the training images via a bootstrapping process, and then use these now-accurately segmented images as a proxy ground-truth in a more standard supervised setting. The key driver for our work is that in the target dataset we typically have reliable ground-truth image-level labels, while data crawled from the web may have unreliable labels, but can be filtered to comprise only easy images to segment, therefore having reliable boundaries. These two forms of information are complementary and we use this observation to build a novel bi-directional transfer learning framework. This framework transfers knowledge between two domains, target domain and web domain, bootstrapping the performance of weakly supervised semantic segmentation. Conducting experiments on the popular benchmark dataset PASCAL VOC 2012 based on both a VGG16 network and on ResNet50, we reach state-of-the-art performance with scores of 60.2% IoU and 63.9% IoU respectively. |
| first_indexed | 2024-10-01T07:02:14Z |
| format | Conference Paper |
| id | ntu-10356/142802 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T07:02:14Z |
| publishDate | 2020 |
| record_format | dspace |
| spelling | ntu-10356/1428022020-07-01T06:45:33Z Bootstrapping the performance of webly supervised semantic segmentation Shen, Tong Lin, Guosheng Shen, Chunhua Reid, Ian School of Computer Science and Engineering 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Engineering::Computer science and engineering Training Image Segmentation Fully supervised methods for semantic segmentation require pixel-level class masks to train, the creation of which is expensive in terms of manual labour and time. In this work, we focus on weak supervision, developing a method for training a high-quality pixel-level classifier for semantic segmentation, using only image-level class labels as the provided ground-truth. Our method is formulated as a two-stage approach in which we first aim to create accurate pixel-level masks for the training images via a bootstrapping process, and then use these now-accurately segmented images as a proxy ground-truth in a more standard supervised setting. The key driver for our work is that in the target dataset we typically have reliable ground-truth image-level labels, while data crawled from the web may have unreliable labels, but can be filtered to comprise only easy images to segment, therefore having reliable boundaries. These two forms of information are complementary and we use this observation to build a novel bi-directional transfer learning framework. This framework transfers knowledge between two domains, target domain and web domain, bootstrapping the performance of weakly supervised semantic segmentation. Conducting experiments on the popular benchmark dataset PASCAL VOC 2012 based on both a VGG16 network and on ResNet50, we reach state-of-the-art performance with scores of 60.2% IoU and 63.9% IoU respectively. Accepted version 2020-07-01T06:45:33Z 2020-07-01T06:45:33Z 2018 Conference Paper Shen, T., Lin, G., Shen, C., & Reid, I. (2018). Bootstrapping the performance of webly supervised semantic segmentation. Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 1363-1371. doi:10.1109/CVPR.2018.00148 978-1-5386-6421-6 https://hdl.handle.net/10356/142802 10.1109/CVPR.2018.00148 2-s2.0-85062847212 1363 1371 en © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/CVPR.2018.00148. application/pdf |
| spellingShingle | Engineering::Computer science and engineering Training Image Segmentation Shen, Tong Lin, Guosheng Shen, Chunhua Reid, Ian Bootstrapping the performance of webly supervised semantic segmentation |
| title | Bootstrapping the performance of webly supervised semantic segmentation |
| title_full | Bootstrapping the performance of webly supervised semantic segmentation |
| title_fullStr | Bootstrapping the performance of webly supervised semantic segmentation |
| title_full_unstemmed | Bootstrapping the performance of webly supervised semantic segmentation |
| title_short | Bootstrapping the performance of webly supervised semantic segmentation |
| title_sort | bootstrapping the performance of webly supervised semantic segmentation |
| topic | Engineering::Computer science and engineering Training Image Segmentation |
| url | https://hdl.handle.net/10356/142802 |
| work_keys_str_mv | AT shentong bootstrappingtheperformanceofweblysupervisedsemanticsegmentation AT linguosheng bootstrappingtheperformanceofweblysupervisedsemanticsegmentation AT shenchunhua bootstrappingtheperformanceofweblysupervisedsemanticsegmentation AT reidian bootstrappingtheperformanceofweblysupervisedsemanticsegmentation |