일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | ||
6 | 7 | 8 | 9 | 10 | 11 | 12 |
13 | 14 | 15 | 16 | 17 | 18 | 19 |
20 | 21 | 22 | 23 | 24 | 25 | 26 |
27 | 28 | 29 | 30 |
Tags
- WSSS
- ssh
- GIT
- 히비스서커스
- vscode
- HookNet
- 티스토리챌린지
- 오블완
- logistic regression
- IVI
- 사회조사분석사2급
- 기초확률론
- 프로그래머스
- CellPin
- docker exec
- docker attach
- AIFFEL
- 백신후원
- Jupyter notebook
- Multi-Resolution Networks for Semantic Segmentation in Whole Slide Images
- Pull Request
- cs231n
- numpy
- cocre
- aiffel exploration
- 도커
- 코크리
- Decision Boundary
- airflow
- docker
Archives
- Today
- Total
목록Region-Based Active Learning for Efficient Labeling in Semantic Segmentation (1)
히비스서커스의 블로그

이 글은 2019년 IEEE 저널에 실린 Region-Based Active Learning for Efficient Labeling in Semantic Segmentation 논문을 읽고 정리한 글입니다. 0. Abstract annotated dataset의 필요성은 증가하고 있으나 모든 픽셀을 annotation 하는 것은 방대하기에 비용이 많이 듦 ⇒ semantic segmentation task에서 region-based active learning 방법을 제안 train dataset에 대해 각 이미지의 10%만 annotation 하여서 train dataset를 annotation한 모델의 93.8%의 성능을 보임 transfer learning을 적용할 때에도 잘 작동하였음 (City..
Theory/Computer Vision
2022. 8. 22. 23:38