[DLHacks]Introduction to ChainerCV
1 like639 views
ChainerCV is a computer vision library built on top of the Chainer deep learning framework. It provides implementations of popular models for feature extraction, object detection, semantic segmentation, and instance segmentation. ChainerCV aims to reproduce the performance of models from original papers and standardizes APIs across models. It also includes useful functions, example codes, documentation, and tests to make it easy to install, use, and maintain.
[DLHacks]Introduction to ChainerCV
- 1. LT: Introduction to ChainerCV Shingo Kitagawa
- 2. What is ChainerCV? • Deep learning computer vision library with Chainer. • ChainerCV is developed by Preferred Networks. • Feel free to make PRs and issues • We are always welcome review and merge. • Implementations • Feature extraction • Object detection • Semantic segmentation • Instance segmentation (New) • and more.. • Github: https://github.com/chainer/chainercv • Arxiv: https://arxiv.org/abs/1708.08169
- 3. Feature extraction • ResNet • Arch: He(original) and Facebook implementation is supported. • ResNet50 • ResNet101 • ResNet152 • VGG • VGG16 • https://chainercv.readthedocs.io/en/stable/reference/links/resnet.html • https://chainercv.readthedocs.io/en/stable/reference/links/vgg.html
- 4. Object detection • Faster-RCNN • Faster-RCNN + VGG16 • Faster-RCNN + ResNet101 • (working) • SSD • SSD300 + VGG16 • SSD512 + VGG16 • YOLO • YOLO v2 + Darknet19 • YOLO v3 + Darknet53 • Training is now working
- 5. Semantic segmentation • SegNet • PSPNet • Training is now working Instance segmentation • FCIS • FCIS + ResNet101 • Training is now working • Mask-RCNN • It will be implemented soon.
- 6. What is different from other CV library? • Reproduce the score of original papers. • Standardize API. • Implement various useful functions. • Prepared sample example code. • Read a lot of WEB documentation. • Easy to install and use. • A lot of test codes for maintenance.
- 7. Reproduction of the original score • Stable: Faster-RCNN, SSD, YOLO, SegNet • No less than 0.5 point lower or even higher • Use it for the comparison with your method! • Experimental: FCIS, PSPNet (New) • Implemented but score is around 1.0 point lower • But we are still working to reproduce original paper.
- 8. API standardization • Models have the same utility functions. • prepare() : Preprocess input • predict() : Predict and return output • use_preset() : Set parameters • Naming conventions • It is strictly fixed in the source code.
- 9. Useful functions • Transforms • crop, flip, resize, translate, pca_lighting, scale, sized_crop • Random + crop, expand, flip, rotate, sized_crop • bounding box, image, point • Visualizations (matplotlib) • vis_bbox, vis_image, vis_point • vis_instance_segmentation, vis_semantic_segmentation • Utils • bbox_iou, mask_iou, nms, I/O • Evaluations • COCO, VOC and more.
- 10. Prepared example codes • chainercv/examples • Demo: • python demo.py • Training: • python train.py • Evaluation: • python eval_voc.py
- 11. WEB documentation (sphinx) • DOC: https://chainercv.readthedocs.io/en/stable/
- 12. Easy to install • pip install chainercv • Additional requirements • ChainerMN, Matplotlib, OpenCV, SciPy • Datasets and models will be installed automatically. • Default: ~/.chainer/pfnet/chainercv • Of course, you can install it manually by yourself. • ROS integration • Rosdistro: python-chainer-pip, python-chainercv-pip
- 13. Test codes • Github + Travis CI • chainercv/tests
- 14. Other Chainer family and useful pages • Chainer is super awesome! • Chainer family • ChainerMN, ChainerRL, ChainerUI • ONNX: ONNX-Chainer • Chainer Research • https://github.com/pfnet-research • chainer-chemistry, sngan, picking-instruction • Chainer community • https://github.com/chainer-community/awesome-chainer • https://github.com/chainer-community/chainer-info • Forum / Slack / Twitter
- 15. If you still cannot love chainer , • PyTorch: kuangliu/torchcv • Inspired from chainercv • Easy to use, read and write • https://github.com/kuangliu/torchcv • Tensorflow: tensorflow/models • A lot of model • Good to read and study, but not good to use. • Caffe2: facebookresearch/Detectron • Easy to use, read and write • Caffe2 is hard to install • Mxnet: apache/incubator-mxnet • A lot of model, but no proof of scores. • Spaghetti codes!