https://github.com/recolude/landmark-recordings

A demo for building recordings from landmark information provided by media pipe
https://github.com/recolude/landmark-recordings

3d computer-vision

Last synced: about 1 year ago
JSON representation

A demo for building recordings from landmark information provided by media pipe

• Host: GitHub
• URL: https://github.com/recolude/landmark-recordings
• Owner: recolude
• Created: 2021-12-22T18:11:08.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2022-01-16T22:20:06.000Z (over 4 years ago)
• Last Synced: 2025-01-23T04:41:25.583Z (over 1 year ago)
• Topics: 3d, computer-vision
• Language: Go
• Homepage:
• Size: 22.5 KB
• Stars: 4
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Landmark Recordings

A demo for building recordings from landmark information provided by [mediapipe](https://mediapipe.dev/).

## Requirements

* FFMPEG

* python 3.9 64-bit (mediapipe requires 64bit)

* Golang 1.17 (I'm sure most versions would work, this is just whats on my machine.)

## Processessing Recordings

### Deconstruct the Video

```bash

ffmpeg -i in.mp4 frames/frame_%04d.png -hide_banner

```

### Run the Landmark Identification Software

```bash

python pose/pose.py

```

### Rebuilding the Video

```bash

ffmpeg -y -r 29 -i frames_out/frame_%04d.png -c:v libx264 -vf fps=29 -pix_fmt yuv420p out.mp4

```

### Buiding a Recolude Recording

```bash

go run pose.go

```