best image labeling tools for Computer Vision tasks:
Conclusion
We've covered four different annotation tools and considered each tool’s advantages and disadvantages:
- labelme can be used for various computer vision tasks, but it involves only manual labeling. However, the tool can be installed and configured very quickly. The tool may be suitable for those who want to annotate a small dataset. : Small dataset에 적합하다.
- labelImg is fit for tasks where only one user will be engaged in labeling for an object detection task. It is also very fast to install and use. Also, it supports various well-known annotation formats.
- CVAT is appropriate for those who work in a team and want to utilize their model to automate the labeling process.
- hasty.ai is not a free tool, but its advantage is that it has built-in assistants that allow you to generate suggested labels after manually annotating 10 images in the dataset, which makes image labeling easier and faster.
출처: https://dida.do/blog/the-best-labeling-tools-for-computer-vision
'DeepLearning' 카테고리의 다른 글
| [경량화] sparse training via boosting pruning plasticity with Neuroregeneration 리뷰 (0) | 2022.01.26 |
|---|---|
| [경량화] Pruning 프루닝 (0) | 2022.01.20 |
| Deep Learning - CH3.1 Probability and Information Theory (0) | 2021.10.24 |
| Naive Bayesian Classifier (0) | 2021.10.15 |
| Prior Probability, Posterior Probability (0) | 2021.10.14 |