a curated list of awesome libraries/ packages/ tools for doing computer vision work on videos with ❤️
Other video related curated lists:
- ffmpeg cheatsheet: a gist of common video operations
- VidGear: a high-performance cross-platform Video Processing framework
- decord: faster video loading experience intended for loading training data for deep learning (medium post)
- imageio: a lightweight package (requires only numpy and pillow) for loading video as an image generator
- ffmpeg-python: one of the most popular ffmpeg python wrapper (examples)
- PyNVVL: a python wrapper of NVIDIA video loader (NVVL) with CuPy for fast video loading with Python
- PySceneDetect: OpenCV based scene cut/ transition detection library with support for both content-based and luminance-based detection
- katna: a color-based keyframes extractor using K-Means (medium post)
- scikit-video: a Python module for video processing built on top of scipy, numpy, and ffmpeg/libav.
- Keyframe Extraction from Videos: simple Jupyter Notebook using Color Histogram, SVD, and Dynamic Clustering method
- VideoFlow: Python framework that facilitates the quick development of complex video analysis applications and other series-processing based applications in a multiprocessing environment (object tracking sample application)
- Video Captioning: combining audio, image captioning, and sentence generation sub-models for video captioning
- Video Analyst: implementation of Single Object Tracking and Video Object Segmentation
- Two-Stream Video Action Recognition Framework: PyTorch implementation of the paper "Towards Good Practices for Very Deep Two-Stream ConvNets"
- Abnormal Event Detection: Keras implementation of "Abnormal Event Detection in Videos using Spatiotemporal Autoencoder"
- ActionVLAD: TensorFlow implementation of "Learning spatio-temporal aggregation for action classification "
- Terran Timeline: a demo using Terran (face recog + pose estimation library) on video that created face timelines for any youtube video
- Video2x: a video upscaler based on waifu2x, a super-res for anime-style art using CNN (youtube demo)
- BasicSR: image and video super-res toolbox based on PyTorch
- Zooming-Slow-Mo: a joint video frame interpolation (VFI) and video super-resolution (VSR) framework built on PyTorch
- video-to-ascii: a python cli to play videos in the terminal using characters as pixels
- pysub2: a python library for editing video subtitle files with support for Advanced SubStation (
.ass
) files (examples) - webvtt-py: read, write, and segment WebVTT caption files in Python