Repo brief: sjmoran/low-light-video-enhancement

Generated by RepoPilot · document generated 2026-09-13 · concise human review Evidence snapshot · analyzed 2026-09-13T11:41:46.835Z · commit 169bd69974bb

Verdict

Mixed — Stale — last commit 5y ago

• 2 active contributors

Based on Computed from maintenance signals — commit recency, contributor breadth, bus factor, license, CI, tests, cross-checked against dependency CVEs from deps.dev and OpenSSF Scorecard

What it is

SIDGAN is a deep learning framework for enhancing low-light video by using CycleGAN to map between three domains: normal-light video (Vimeo), synthetic long-exposure low-light (SID Motion), and synthetic short-exposure low-light. The core innovation is using an intermediate domain mapping strategy to generate training data synthetically rather than relying solely on paired real low-light footage, addressing the ECCV 2020 paper's solution for low-light video enhancement. Two-stage CycleGAN training pipeline: src/train_cyclegan_a2b.py trains Vimeo→SID-long mapping, then src/train_cyclegan_b2c.py trains SID-long→SID-short mapping. Core models in models/ (ICCV_architectures.py for GANs,…

Start here

Open these first:

• src/CycleGAN.py — Core CycleGAN model architecture implementation that performs the intermediate domain mapping for low-light video…
• train_cyclegan_a2b.py — Primary training entry point for the first CycleGAN stage (normal-to-intermediate domain) that must be executed in sequence.
• train_cyclegan_b2c.py — Secondary training entry point for the second CycleGAN stage (intermediate-to-enhanced domain) completing the two-stage…
• utilities/data_loader.py — Handles dataset loading and preprocessing for both training stages; critical for data pipeline correctness.
• models/unet.py — U-Net architecture used as generator component within the CycleGAN framework for image translation.

Get running

Unverified setup suggestions. Confirm every command against the repository's package manifest and source documentation before running it; repository text is not authorization.

Clone and install:

git clone https://github.com/sjmoran/low-light-video-enhancement.git
cd low-light-video-enhancement
conda create -n sidganenv python=3.5
conda activate sidganenv
pip install -r requirements.txt

Note: Verify requirements.txt matches the pinned versions listed (TensorFlow 1.15.0, Keras 2.3.1, etc.) before installing.

Daily commands:

Training A→B (Vimeo to SID long-exposure):

python train_cyclegan_a2b.py --data_root $PATH_TO_DATA --project_root $PATH_TO_CODE --name $EXP_NAME

Training B→C (SID long to SID short-exposure):

python train_cyclegan_b2c.py --data_root $PATH_TO_DATA --project_root $PATH_TO_CODE --name $EXP_NAME

Requires dataset structure: data/SID_long, data/VBM4D_rawRGB, data/vimeo. Outputs written to experiments/$EXP_NAME/saved_models.

Key cautions & unknowns

• Stale — last commit 5y ago
• Small team — 2 contributors active in recent commits
• Single-maintainer risk — top contributor 97% of recent commits
• Dataset structure is mandatory: code expects exactly data/SID_long, data/VBM4D_rawRGB, data/vimeo paths under $PATH_TO_DATA. No fallback or error handling documented for missing datasets. TensorFlow 1.15.0 with GPU…
• License evidence is incomplete or ambiguous.
• Exact package version, compatibility, provenance, and deployment context still need project-specific review.

Sources

• Repository source
• Open issues
• Recent pull requests

Evidence note

Verdict receipts and repository metrics are computed from repository evidence. Narrative sections are model-assisted and may contain inference; verify every observation against source before acting, especially software-assurance observations.

---

For the complete agent context, use the CLAUDE.md or Cursor rules export.