lyrics_overlay

What?

This project aims to put lyrics for a particular frame of a video in an optimal location avoiding overlap with the main subjects of the video.

Assumptions:

1. In this case "subjects" are "persons".
2. Number of subjects = 1
3. Maximum time-length = 5 minutes
4. Maximum frames per second = 60
5. Maximum resolution of each frame (or video) = 1920x1080 (FHD)
6. No time-overlapping lyrics

Architecture

How to run?

Pre-requisites

1. Since its a Python project - one needs to have Python v3.8+.
2. Clone the repo.
3. Make and activate a virtual environment either from your IDE or by following the instructions here.
4. Install all the pip dependencies by running - pip3 install -r requirements.txt.

For Mac users

The torch and torchvision can be installed by running pip3 install torch torchvision.

Running the application

Run this command inside your virtual environment - python3 main.py

For developing

This repo is auto-linted using pre-commit. So before you push any commits please install it following the instructions in their official website here.