In this project, we applied yolov7 to localize bounding box of boxes and segment them. To illustrate this, see figure below.
- Table of Contents
- Install Requirements
- Create and activate virtual environment
- Data Format
- Generate auto-augmented data
- Download Yolov7 Weights
- Evaluation
- Training
- Result
- Citation
- Acknowledgement
Expand
| Package | Version |
|---|---|
| absl-py | 1.4.0 |
| albumentations | 1.3.0 |
| appdirs | 1.4.4 |
| attrs | 22.2.0 |
| backcall | 0.2.0 |
| cachetools | 5.3.0 |
| certifi | 2022.12.7 |
| chardet | 4.0.0 |
| charset-normalizer | 3.1.0 |
| clearml | 1.10.3 |
| click | 8.1.3 |
| clip | 0.2.0 |
| cycler | 0.11.0 |
| decorator | 5.1.1 |
| docker-pycreds | 0.4.0 |
| fonttools | 4.38.0 |
| furl | 2.1.3 |
| gitdb | 4.0.10 |
| GitPython | 3.1.31 |
| google-auth | 2.17.2 |
| google-auth-oauthlib | 0.4.6 |
| grpcio | 1.53.0 |
| idna | 3.4 |
| imageio | 2.27.0 |
| imgaug | 0.4.0 |
| importlib-metadata | 6.3.0 |
| importlib-resources | 5.12.0 |
| ipython | 7.34.0 |
| jedi | 0.18.2 |
| joblib | 1.2.0 |
| jsonschema | 4.17.3 |
| kiwisolver | 1.4.4 |
| Markdown | 3.4.3 |
| MarkupSafe | 2.1.2 |
| matplotlib | 3.5.3 |
| matplotlib-inline | 0.1.6 |
| networkx | 2.6.3 |
| numpy | 1.21.6 |
| nvidia-cublas-cu11 | 11.10.3.66 |
| nvidia-cuda-nvrtc-cu11 | 11.7.99 |
| nvidia-cuda-runtime-cu11 | 11.7.99 |
| nvidia-cudnn-cu11 | 8.5.0.96 |
| oauthlib | 3.2.2 |
| opencv-python | 4.7.0.72 |
| opencv-python-headless | 4.7.0.72 |
| orderedmultidict | 1.0.1 |
| packaging | 23.0 |
| pandas | 1.3.5 |
| parso | 0.8.3 |
| pathlib2 | 2.3.7.post1 |
| pathtools | 0.1.2 |
| pexpect | 4.8.0 |
| pickleshare | 0.7.5 |
| Pillow | 9.5.0 |
| pip | 23.0.1 |
| pkgutil_resolve_name | 1.3.10 |
| prompt-toolkit | 3.0.38 |
| protobuf | 3.20.1 |
| psutil | 5.9.4 |
| ptyprocess | 0.7.0 |
| pyasn1 | 0.4.8 |
| pyasn1-modules | 0.2.8 |
| pycocotools | 2.0.6 |
| Pygments | 2.14.0 |
| PyJWT | 2.4.0 |
| pyparsing | 3.0.9 |
| pyrsistent | 0.19.3 |
| python-dateutil | 2.8.2 |
| python-dotenv | 0.21.1 |
| pytz | 2023.3 |
| PyWavelets | 1.3.0 |
| PyYAML | 6.0 |
| qudida | 0.0.4 |
| requests | 2.28.2 |
| requests-oauthlib | 1.3.1 |
| requests-toolbelt | 0.10.1 |
| roboflow | 1.0.3 |
| rsa | 4.9 |
| scikit-image | 0.19.3 |
| scikit-learn | 1.0.2 |
| scipy | 1.7.3 |
| seaborn | 0.12.2 |
| sentry-sdk | 1.19.1 |
| setproctitle | 1.3.2 |
| setuptools | 47.1.0 |
| shapely | 2.0.1 |
| six | 1.16.0 |
| smmap | 5.0.0 |
| tensorboard | 2.11.2 |
| tensorboard-data-server | 0.6.1 |
| tensorboard-plugin-wit | 1.8.1 |
| thop | 0.1.1.post2209072238 |
| threadpoolctl | 3.1.0 |
| tifffile | 2021.11.2 |
| torch | 1.10.1+cu102 |
| torch-tb-profiler | 0.4.1 |
| torchaudio | 0.10.1+cu102 |
| torchvision | 0.11.2+cu102 |
| tqdm | 4.65.0 |
| traitlets | 5.9.0 |
| typing_extensions | 4.5.0 |
| ultralytics | 8.0.110 |
| urllib3 | 1.26.15 |
| wandb | 0.14.2 |
| wcwidth | 0.2.6 |
| Werkzeug | 2.2.3 |
| wget | 3.2 |
| wheel | 0.40.0 |
| zipp | 3.15.0 |
Expand
# Create virtual environment
python3 -m venv venv_yolo
# Activate virtual environment
source path/to/ven_yolo/bin/activate
Expand
Yolov7_box_segmentation
|
|
|____Box_data
| |
| |________Filling_2023_05_16
| |
| |______Test
| | |________.png/.jpg
| |
| |
| |______Train
| | |
| | |_________.bmg
| | |
| | |_________.json
| |
| |_______Val # Test image
| |
| |_________.bmg
| |
| |_________.json
|
|
|______command
| |
| |_________predict_UK_OK.sh
| |
| |_________predict.sh
| |
| |_________train.sh
|
|______seg
|
|_____data
| |
| |_______hyps
| | |_______blister.yaml
| | |_______coco.yaml
| | |_______hyp.scratch.custom.yaml
| | |_______hyp.scratch.p5.yaml
| |
| |_______scripts
| | |
| | |_______get_coco.sh
| | |
| | |_______get_imagenet.sh
| |
| |
| |_______coco.yaml
|
|
|______models
|
|______runs
|
|______segment
|
|______tools
|
|______utils
|
|______wandb
|
|______export.py # run inference file
|
|______hubconf.py
|
|______requirements.txt # Evaluate on new data
|
|______yolov5m-seg.pt # Test
|
|______yolov5s-seg.pt # Train
Expand
# Convert polygon read from json file to mask png/jpg file
python3 cvtpoly2_mask.py # (this code will save mask in folder ./result_mask) (1)
# Read background image and mask.jpg created from script (1) to replace background
python3 replace_background.py # (this script will replace image with background) (2)
# Read background image and mask.jpg created from script (1) to replace background
python3 convert_data.py # (this script will convert images, label, masks to coco format to train) (3)
yolov7.pt yolov7x.pt yolov7-w6.pt yolov7-e6.pt yolov7-d6.pt yolov7-e6e.pt
# Infer boxes and their segments
python3 seg/segment/predict.py
--weights=/home/tonyhuy/Yolov7_box_segmentation/seg/runs/train-seg/exp5/weights/epoch_111_best.pt
--source=/home/tonyhuy/Yolov7_box_segmentation/Box_data/bad
--data=/home/tonyhuy/Yolov7_box_segmentation/seg/data/coco.yaml
--imgsz=320
--conf-thres=0.8
--device=2
Run this script to generate augmented images.
# If save augmented data,use this command
#____________________________________________ Run training file___________________________________________________#
# Image size 320
python3 segment/train.py
--weights=/home/tonyhuy/Yolov7_box_segmentation/seg/yolov5s-seg.pt
--data=/home/tonyhuy/Yolov7_box_segmentation/seg/data/coco.yaml
--hyp=/home/tonyhuy/Yolov7_box_segmentation/seg/data/hyps/experiment.yaml
--epochs=300
--img=320
--device=1
# Image size 640
python3 segment/train.py
--weights=/home/tonyhuy/Yolov7_box_segmentation/seg/yolov5s-seg.pt
--data=/home/tonyhuy/Yolov7_box_segmentation/seg/data/coco.yaml
--hyp=/home/tonyhuy/Yolov7_box_segmentation/seg/data/hyps/experiment.yaml
--epochs=300
--img=640
--device=1
WANDB
Weights and Bias running result
Prediction result
@article{wang2022yolov7,
title={{YOLOv7}: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors},
author={Wang, Chien-Yao and Bochkovskiy, Alexey and Liao, Hong-Yuan Mark},
journal={arXiv preprint arXiv:2207.02696},
year={2022}
}
@article{wang2022designing,
title={Designing Network Design Strategies Through Gradient Path Analysis},
author={Wang, Chien-Yao and Liao, Hong-Yuan Mark and Yeh, I-Hau},
journal={arXiv preprint arXiv:2211.04800},
year={2022}
}
Expand
- https://github.com/WongKinYiu/yolov7
- https://github.com/AlexeyAB/darknet
- https://github.com/WongKinYiu/yolor
- https://github.com/WongKinYiu/PyTorch_YOLOv4
- https://github.com/WongKinYiu/ScaledYOLOv4
- https://github.com/Megvii-BaseDetection/YOLOX
- https://github.com/ultralytics/yolov3
- https://github.com/ultralytics/yolov5
- https://github.com/DingXiaoH/RepVGG
- https://github.com/JUGGHM/OREPA_CVPR2022
- https://github.com/TexasInstruments/edgeai-yolov5/tree/yolo-pose