Uni-Evaluator: A unified interactive model evaluation for classification, object detection, and instance segmentation in computer vision
Uni-Evaluator is a visual analysis tool to support a unified model evaluation for different computer vision tasks, including classification, object detection, and instance segmentation. More information, including the video and live demo, can be found here.
We further cleaned the annotations of the COCO validation dataset after the case study shown in our paper. The COCO-format annotation file can be downloaded via this link.
git clone https://github.com/thu-vis/Uni-Evaluator.git
cd Uni-Evaluator(This repo is tested with node_v16 and python3.8 on Ubuntu)
npm install -g yarn
cd frontend
yarn# install from requirements.txt
cd ../backend
# install the packages in order
cat requirements.txt | xargs pip install
# install RangeTree
cd ./data/RangeQuery
python setup.py build_ext --inplace
cd ../../
# install fastlapjv
git clone [email protected]:thu-vis/fast-lapjv.git
cd fast-lapjv/
python setup.py install --user
cd ../
# install faiss (you can also follow https://github.com/facebookresearch/faiss/blob/main/INSTALL.md)
conda install -c pytorch faiss-gpu
# or
conda install -c pytorch faiss-cpuWe provide COCO test data used in our case study for demonstration, which can be downloaded here. It corresponds to the dataPath below.
# start backend
cd backend
python server.py \
--dataPath <path> \ # contains sub-directories with training or test data
--seg \ # added for segmentation task
--dataName <name> # the name of dataset, e.g., "COCO", "iSAID"
--host <host> \
--port <port> \
cd ../
# start frontend
cd frontend
yarn startIf you use this code or the cleaned annotations of COCO validation data for your research, please consider citing:
@article{chen2023unified,
title={A unified interactive model evaluation for classification, object detection, and instance segmentation in computer vision},
author={Chen, Changjian and Guo, Yukai and Tian, Fengyuan and Liu, Shilong and Yang, Weikai and Wang, Zhaowei and Wu, Jing and Su, Hang and Pfister, Hanspeter and Liu, Shixia},
journal={IEEE Transactions on Visualization and Computer Graphics},
year={2023},
publisher={IEEE}
}
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent