# soCzech/TransNetV2

TransNet V2: Shot Boundary Detection Neural Network

Repository: https://github.com/soCzech/TransNetV2
Canonical: https://ross.abutalabs.com/products/transnetv2
Homepage: https://arxiv.org/abs/2008.04838
Language: Python
License: MIT
License Family: permissive
Topics: shot-boundary-detection, shot-detection
Last push: 2023-12-04T20:06:15+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2509, "days_push": 1003, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1027, forks 149 (observed 2026-08-28T04:03:17.184156+00:00)

## What it is
TransNet V2 is a deep neural network for shot boundary detection in videos, achieving state-of-the-art results on benchmarks like ClipShots and BBC Planet Earth. The repository provides pretrained models for inference (TensorFlow and PyTorch) plus training code to replicate the paper's experiments.

## Use cases
- detect shot transitions in a video
- split a movie into scenes automatically
- find hard cuts and gradual transitions in video files
- segment video archives into shots for indexing
- run pretrained shot detection model in python
- compare shot boundary detection methods

## When to choose
- you need accurate, fast shot boundary detection with a pretrained model
- you want state-of-the-art F1 scores on standard shot detection benchmarks
- you need both TensorFlow and PyTorch inference options
- you want to reproduce or extend published shot detection research

## When to avoid
- you need general video editing or transcoding rather than shot detection
- you want a maintained product with frequent updates and support
- you cannot use GPU acceleration for inference
- you need audio-based scene segmentation

## Facets
- artifact type: library
- maturity: stable
- function: computer-vision, video-processing, deep-learning, machine-learning
- domain: computer-vision, machine-learning, deep-learning
- platform: python
- tags: shot-boundary-detection, scene-detection, video-segmentation, pytorch, tensorflow, pretrained-model, research-code, video, gpu, docker, linux

## Member repositories
- soCzech/TransNetV2 (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:17.184156+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:07:55.088397+00:00, confidence not recorded.
  - readme: https://github.com/soCzech/TransNetV2 (fetched 2026-08-28T04:03:17.184156+00:00, sha 7dd99a163097)
  - homepage: https://arxiv.org/abs/2008.04838 (fetched 2026-08-29T13:08:23.743811+00:00, sha fee4826b76c1)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T13:08:23.753128+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T13:08:23.756631+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T13:08:23.758663+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T13:08:23.754997+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.