Face Timelapse Generator

A Python tool that automatically detects and aligns faces across a series of photos to produce a smooth, face-centered timelapse video.

How It Works

Given a folder of images (e.g. selfies taken over months or years), the script:

1. Detects the largest face in each image using OpenCV Haar cascades
2. Scales and centers the face to a fixed position in every frame
3. Optionally corrects head rotation using eye detection
4. Writes all frames to an .avi video file in the correct chronological order

Demo

https://youtu.be/efOw1FrIRHU

Requirements

opencv-python>=4
numpy

Install with:

pip install -r requirements.txt

Usage

1. Put your images (.jpg) in the ./images/ folder, named numerically (1.jpg, 2.jpg, …)
2. Run the script:

python face_lapse.py

If a face cannot be automatically detected in an image, it will be skipped. Set manual_mode = True in main() to manually draw a bounding box around the face instead.

Configuration

Inside face_lapse.py, the main() function exposes these parameters:

Parameter	Default	Description
fps	8.0	Frames per second in the output video
face_height	400	Target face height in pixels
video_out_size	(1280, 720)	Output video resolution
manual_mode	False	Fall back to manual face selection on detection failure
correct_colour	False	Experimental CLAHE colour correction
min_haar_face_size	(200, 200)	Minimum face size for Haar detection

Output

The output video is saved as video.avi in the project root.

Project Structure

face_pro_lapse/
├── images/                  # Input images (1.jpg … N.jpg)
├── face_lapse.py            # Main script
├── rename_images.py         # Utility: clean up and reorder image filenames
├── haarcascade_*.xml        # OpenCV Haar cascade classifiers
├── requirements.txt
└── video.avi                # Generated timelapse output