The YouTube Object Detection is a Python-based application that aims to process YouTube videos and detect objects within the video frames using computer vision techniques. The project utilizes state-of-the-art object detection algorithms and libraries to automatically identify and localize objects of interest in the video frames.
OpenCV
OpenCV (Open Source Computer Vision Library) is a popular open-source computer vision and machine learning library for image and video processing. Written in C++ and Python, it offers a wide range of tools for tasks like image manipulation, object detection, and camera calibration. Widely used in various fields, OpenCV is known for its comprehensive documentation, large community support, and cross-platform compatibility.
YOLOv4-tiny
YOLO v4 Tiny is a compact object detection algorithm designed for resource-constrained environments. It offers a balance between accuracy and inference speed, with a smaller network architecture optimized for real-time object detection. It can be implemented using popular deep learning frameworks and trained on custom datasets for specific domains, maintaining accuracy despite its smaller size.
FastAPI
FastAPI is a modern, fast, and lightweight Python web framework for building APIs. It offers automatic documentation generation, data validation, and serialization using type hints, making it highly productive. It also supports asynchronous programming and is popular among developers for building scalable APIs with Python.
$ git clone https://github.com/felipeall/youtube-object-detection.git
$ cd youtube-object-detection
$ poetry shell
$ poetry install
$ make run-local
$ open localhost:8080$ git clone https://github.com/felipeall/youtube-object-detection.git
$ cd youtube-object-detection
$ make run
$ open localhost:8080