Using FFMPEG to Encode-Decode Video for Offline Video-Based Machine Learning

Learn how to use FFMPEG for video encoding and decoding in machine learning applications. This 21-minute tutorial demonstrates techniques for processing individual video frames, generating new videos, and converting between formats like MP4/MOV and JPEG images. Explore three Jupyter notebooks covering topics such as generating faces, prototyping seeds, latent vectors, morphing, randomness, and syncing audio. Discover how to modify images, build video files, and implement YOLO object detection. Gain practical insights into file setup, virtual machine configuration, and optimizing output quality for offline video-based machine learning projects.

Introduction
Generating Faces
Prototype Seeds
Latent vectors
Morph step
Randomness
Running FFMPEG
Results
Putting Images Together
Basic Example
Uploading Longer Videos
Execute Command
Run FFMPEG
Output
Input
Output Quality
Extracting Audio
Syncing Audio
Running Process
Demonstration
Ball Speed
File Setup
Modifying Images
Building the Video File
File Names
File Download
Demo
Modifications
Yolo
Collab
Fixing Hardcoded Video
Yolo Eve
Tensorflow
Virtual Machine
Yellow Waits
Set Up Yellow
Time Formatter
Process Image
Fast Forward
Build Video File
Final Output