#AI-Shorts-Creator! ๐ฅโ๏ธ (WIP, You Might face some Bugs)
AI-Shorts-Creator is a powerful tool designed for content creators, podcasters, and video enthusiasts to effortlessly extract captivating segments from their videos. Leveraging the advanced language model GPT-4, this innovative solution intelligently analyzes video transcripts to identify the most viral and exciting moments. By harnessing the capabilities of FFmpeg and OpenCV, AI-Shorts-Creator automatically crops videos, allowing you to focus on the key highlights and provide an enhanced viewing experience.
- Automatically extract captivating segments from videos.
- Identify the most viral and exciting moments using GPT-4.
- Crop videos to emphasize key highlights with precise face detection.
- Streamline video editing and save time by eliminating manual searching.
- Work seamlessly with various video formats for maximum compatibility.
- Enhance the viewing experience for your audience with perfectly cropped highlights.
Source Video : https://www.youtube.com/watch?v=NHaczOsMQ20
output_shorts.mp4
Get started with AI-Shorts-Creator today and unlock the potential of your videos like never before!
Requirements
- Python 3.x
pytube
library (install withpip install pytube
)opencv-python
library (install withpip install opencv-python
)openai
library (install withpip install openai
)youtube-transcript-api
library (install withpip install youtube-transcript-api
)- FFmpeg (install according to your operating system)
- Install the required libraries by running the following command:
pip install pytube opencv-python openai youtube-transcript-api
-
Install FFmpeg by following the installation instructions for your operating system. Make sure the
ffmpeg
command is accessible from the command line. -
Set up your OpenAI API key by replacing
openai.api_key = ''
with your actual OpenAI API key. -
Modify the
video_id
variable in themain()
function to specify the YouTube video you want to process. -
Run the script:
python auto_cropper.py
The script will download the YouTube video, analyze its transcript using OpenAI's GPT-4, extract the best sections based on the analysis, crop the video using FFmpeg, and apply face detection using OpenCV to further refine the cropping.
- The
download_video(url, filename)
function downloads a YouTube video by providing the URL and specifying the filename. - The
segment_video(response)
function segments the video into interesting sections based on a transcript analysis using OpenAI's GPT-4 model. - The
detect_faces(video_file)
function uses face detection to identify faces in a video file. - The
crop_video(faces, input_file, output_file)
function crops the video around the detected faces using FFmpeg. - The
is_talking_in_batch(frames)
function analyzes the lip movement or facial muscle activity within a batch of frames to determine if talking behavior is present. - The
adjust_focus(frame, talking)
function applies visual effects or adjustments to emphasize the speaker in the frame.
Please note that the GPT-4 model and transcript analysis functionality in the provided code are simulated and not fully functional. You would need a valid OpenAI API key and a working GPT-4 model to perform transcript analysis.