This repository benchmarks the performance of PySceneDetect in terms of both latency and accuracy. We evaluate it using the standard dataset for video shot detection: BBC and AutoShot.
# annotation
wget -O BBC/fixed.zip https://zenodo.org/records/14873790/files/fixed.zip
unzip BBC/fixed.zip -d BBC
rm -rf BBC/fixed.zip
# videos
wget -O BBC/videos.zip https://zenodo.org/records/14873790/files/videos.zip
unzip BBC/videos.zip -d BBC
rm -rf BBC/videos.zip
Download AutoShot_test.tar.gz
from Google drive.
tar -zxvf AutoShot.tar.gz
rm AutoShot.tar.gz
To evaluate PySceneDetect on a dataset, run the following command from the root of the repo:
python -m benchmark --dataset <dataset_name> --detector <detector_name>
For example, to evaluate ContentDetector on the BBC dataset:
python -m benchmark --dataset BBC --detector detect-content
To run all detectors on all datasets:
python -m benchmark --all
The --all
flag can also be combined with --dataset
or --detector
.
The performance is computed as recall, precision, f1, and elapsed time.
Detector | Recall | Precision | F1 | Elapsed time (second) |
---|---|---|---|---|
AdaptiveDetector | 87.12 | 96.55 | 91.59 | 27.84 |
ContentDetector | 84.70 | 88.77 | 86.69 | 28.20 |
HashDetector | 92.30 | 75.56 | 83.10 | 16.00 |
HistogramDetector | 89.84 | 72.03 | 79.96 | 15.13 |
ThresholdDetector | 0.00 | 0.00 | 0.00 | 18.95 |
Detector | Recall | Precision | F1 | Elapsed time (second) |
---|---|---|---|---|
AdaptiveDetector | 70.77 | 77.65 | 74.05 | 1.23 |
ContentDetector | 63.67 | 76.40 | 69.46 | 1.21 |
HashDetector | 56.66 | 76.35 | 65.05 | 1.16 |
HistogramDetector | 63.36 | 53.34 | 57.92 | 1.23 |
ThresholdDetector | 0.75 | 38.64 | 1.47 | 1.24 |
@InProceedings{bbc_dataset,
author = {Lorenzo Baraldi and Costantino Grana and Rita Cucchiara},
title = {A Deep Siamese Network for Scene Detection in Broadcast Videos},
booktitle = {Proceedings of the 23rd ACM International Conference on Multimedia},
year = {2015},
}
@InProceedings{autoshot_dataset,
author = {Wentao Zhu and Yufang Huang and Xiufeng Xie and Wenxian Liu and Jincan Deng and Debing Zhang and Zhangyang Wang and Ji Liu},
title = {AutoShot: A Short Video Dataset and State-of-the-Art Shot Boundary Detection},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
year = {2023},
}