obi19999/smart-video-reframe

View on GitHub by @obi19999

🎥 Auto-reframe videos for mobile using AI to track faces and analyze scenes, ensuring optimal framing for platforms like TikTok and Reels.

What's novel

🎥 Auto-reframe videos for mobile using AI to track faces and analyze scenes, ensuring optimal framing for platforms like TikTok and Reels.

Code Analysis

9 files read · 4 rounds

A CLI tool that uses YOLO face detection to crop landscape videos into vertical 9:16 format by centering detected faces, with temporal smoothing of bounding boxes and scene-aware segmentation.

Strengths

The temporal smoothing pipeline (EMA + face count homogenization + group size stabilization) shows genuine thought about making face tracking stable across frames. The two-mode design (AI crop vs blur-fit) provides a reasonable fallback for videos without detectable faces.

Weaknesses

The README describes a GUI desktop application with installers that doesn't exist in the code — it's purely a CLI tool. Critical bugs include loading all frames into RAM (OOM on any real video), an undefined variable crash path in the output loop, and zero tests. Debug prints are left in production code and the YOLO model file is missing from the repo.

Score Breakdown

Innovation

4 (25%)

Craft

31 (35%)

Traction

9 (15%)

Scope

30 (25%)

Signal breakdown

Innovation

Not Fork+1

Code Novelty+1

Concept Novelty+1

Craft

Ci-3

Tests-4

Polish+0

Releases+0

Has License+5

Code Quality+9

Readme Quality+15

Recent Activity+7

Structure Quality+5

Commit Consistency+0

Has Dependency Mgmt+5

Traction

Forks+0

Stars+6

Hn Points+0

Watchers+3

Early Traction+0

Devto Reactions+0

Community Contribs+2

Scope

Commits+3

Languages+3

Subsystems+5

Bloat Penalty+0

Completeness+7

Contributors+6

Authored Files+8

Readme Code Match+3

Architecture Depth+5

Implementation Depth+8

Evidence

Commits

Contributors

Files

Active weeks

TestsCI/CDREADMELicenseContributing

Repository

Language

Python

Stars

Forks

License

MIT