roboflow/supervision
Supervision
Supervision is a computer vision toolset for normalizing model outputs, managing datasets, and visualizing annotations. It provides a framework to convert predictions from various classification and detection models into a standardized data format to ensure interoperability across different computer vision pipelines.
The library features a post-processor for filtering, counting, and tracking detected objects across image frames and video streams. It includes capabilities for large image tiling to improve the detection of small objects and tools for assigning persistent identities to objects to monitor movement across polygonal zones and boundary lines.
The project covers a broad range of computer vision utilities, including dataset format conversion, image geometry transformations, and the rendering of visual overlays such as bounding boxes and masks. It also provides tools for benchmarking model performance using standardized accuracy metrics and confusion matrices.
Features
- Computer Vision Post-Processing Toolsets - Offers a comprehensive collection of utilities for processing, normalizing, and visualizing detections from various computer vision models.
- Computer Vision Normalizers - Provides a framework to convert predictions from diverse computer vision models into a standardized data format.
- Bounding Box Visualizers - Draws bounding boxes, masks, and labels on images and video to visually verify model predictions.
- Standardized Detection Formats - Converts imagery processed through various computer vision models into a standardized detection format for interoperability.
