aparsoft/yolo-streamlit-detection-tracking

Object detection and tracking algorithm implemented for Real-Time video streams and static images.

What it solves

This project provides a comprehensive web-based interface for real-time computer vision tasks. It simplifies the process of deploying and testing YOLO-based models for object detection, instance segmentation, pose estimation, and object tracking without requiring complex setup or custom code for different input sources.

How it works

Built with Streamlit, the application acts as a wrapper around the Ultralytics YOLO26 and YOLO World v2 models. It uses a modular service-based architecture to handle different inference modes:

  • Image Inference: Processes uploaded images or defaults, applying the selected vision task.
  • Video Inference: Handles streams from stored files, WebRTC-enabled webcams, RTSP streams, and YouTube URLs. It integrates ByteTrack and BoTSORT for object tracking and implements a frame-skipping mechanism to optimize performance on long videos.
  • Open-Vocabulary Detection: Leverages YOLO World v2 to allow users to detect objects using natural language text prompts instead of fixed class labels.

Who it’s for

  • Developers and Researchers: Those looking to quickly prototype and test vision models on various media types.
  • AI Enthusiasts: Users who want to explore object detection and pose estimation through a GUI rather than writing Python scripts.
  • Deployment Testers: People needing a browser-native way to test webcam feeds and RTSP streams using YOLO models.

Highlights

  • Multi-Task Support: Combines detection, segmentation, and pose estimation in one app.
  • Open-Vocabulary Detection: Supports natural language prompts for flexible object searching.
  • Couting & Tracking: Features both per-frame local and cumulative global object counting.
  • Flexible Video Sources: Native browser webcam support via streamlit-webrtc, RTSP, and YouTube integration.
  • Performance Tuning: Includes a frame-skip slider (1-8x) to accelerate video processing.

Related

  • Project
    ultralytics/ultralytics

    A comprehensive computer vision framework featuring YOLO models for real-time object detection, segmentation, classification, and pose estimation.

  • Project
    RizwanMunawar/yolov7-object-tracking

    A computer vision project that integrates YOLOv7 and YOLOv8 for real-time object detection and tracking across video streams from various sources.

  • Project
    z1069614715/objectdetection_script

    A comprehensive toolkit for improving and compressing object detection models, featuring modified YOLO and RT-DETR architectures, pruning, distillation, and multimodal support.

  • Project
    yzfzzz/depth-detect

    A high-performance C++ and TensorRT framework that fuses YOLO object detection with monocular/binocular depth estimation to track objects and estimate their motion states.

  • Project
    whitphx/streamlit-webrtc

    A Streamlit component for handling real-time video and audio streaming via WebRTC, enabling the creation of interactive media processing apps in Python.