hendkai/cut_youtube_to_word_video

Cut a youtube video down to word videos and collections the information in a sqlite. Later it can be used to build a video just from a text.

🎬 Word-Based Video Processing (not finished)

📝 Description

This project comprises two main scripts: cut_youtube_to_word_video.py and create_video.py. cut_youtube_to_word_video.py is used to download YouTube videos, convert them into audio, transcribe speech to text, and then slice the videos based on individual spoken words. create_video.py allows creating a continuous video from individual word videos based on a given sentence.

🚀 How it Works

📹 cut_youtube_to_word_video.py

• Downloads videos from YouTube.
• Converts videos to audio format.
• Utilizes Google Cloud Speech-to-Text API for transcribing audio recordings.
• Slices the video into segments corresponding to individual words.
• Stores information about the videos in an SQLite database.

🎞️ create_video.py

• Creates a video from a given sentence by combining corresponding word videos.
• Displays a list of available words with video files.
• Allows the user to input a sentence to create a video from.

🔒 License

This project is currently not available under an open-source license. All rights reserved. The use, reproduction, or distribution of the code without express permission of the author is prohibited.

🤖 Developer Note

Portions of this project were developed with assistance from OpenAI's ChatGPT, particularly in scripting and troubleshooting.

📦 Requirements

• Python 3
• PyTube
• MoviePy
• Google Cloud Speech-to-Text API
• SQLite3
• OpenCV
• Tkinter

💻 Installation and Execution

(Here, include the steps for installation and how to run your scripts.)

🛣️ Roadmap

1. Add Video Rating Model: Implement a machine learning model to automatically rate the quality of word videos based on clarity and relevance. The model will also determine the ideal start and end times for each word segment, optimizing the accuracy of video slicing. To achieve this:
   • Collect a dataset of word videos with annotations for quality, clarity, start, and end times.
   • Train a supervised machine learning model, possibly a neural network, to predict these parameters.
   • Integrate the model into the cut_youtube_to_word_video.py script to automatically adjust the slicing of videos.

---

⚠️ Disclaimer: This README is for illustrative purposes. Please adjust the content to fit your actual project's description and requirements.