Open-source alternatives to Supervision

Labelme is a Python-based image annotation tool used to create computer vision datasets. It serves as a visual editor for semantic segmentation, allowing users to define object boundaries using polygons, rectangles, points, and circles. The application also functions as a multispectral image annotator, supporting high-bit depth TIFF files used in satellite and scientific imagery. The tool incorporates AI-assisted labeling capabilities to automate the creation of masks and polygons. These features allow for shape generation driven by text prompts or interactive point selections, which propose

SAHI is a sliced inference framework and computer vision pipeline designed to detect small objects in high-resolution images. It provides a system for dividing large images into overlapping patches to prevent the detail loss that typically occurs during standard model downscaling, alongside an image tiling utility and a COCO dataset toolkit. The project distinguishes itself by offering a model-agnostic prediction wrapper that standardizes different machine learning frameworks into a unified interface. This allows it to implement sliced inference and object detection across various model backe

imutils is a computer vision utility toolkit and image processing library designed to simplify common manipulation tasks using OpenCV. It serves as an image analysis helper and geometry transformation tool for automating visual data processing. The toolkit provides specialized capabilities for maintaining image integrity during transformations, such as resizing images while preserving aspect ratios and rotating images without cropping corners. It also includes tools for four-point perspective warping to create top-down views and the extraction of topological skeletons from binary images. The

MMDetection3D is an open-source toolbox for 3D perception, providing a unified framework for detecting and segmenting objects in three-dimensional environments. It supports a range of core tasks including monocular 3D object detection from single camera images, LiDAR-based 3D object detection from raw point clouds, and multi-modal fusion that combines camera images with LiDAR data. The toolbox also covers point cloud semantic segmentation, assigning class labels to every point in a scan for scene understanding. The project distinguishes itself through a config-driven pipeline that orchestrate

CVAT is an open-source computer vision annotation tool and visual dataset management platform. It provides a self-hosted interface for labeling images, videos, and 3D data to create datasets for vision AI models. The platform features AI-assisted data labeling to automate the creation of masks and bounding boxes, utilizing a plug-in system to connect external machine learning models. It includes a consensus-based quality assurance system that verifies label accuracy by comparing independent annotations. The system covers collaborative team management, project organization through task decomp

tracking.js is a browser computer vision library written in JavaScript for performing real-time image analysis and object tracking directly within a web browser. It functions as a real-time object tracker, a color tracking tool, and a face detection utility. The library enables the detection and monitoring of specific color ranges, human faces, and known visual patterns across consecutive video frames. It extracts visual features and descriptors from images to identify distinct landmarks for matching and tracking. The project covers broad computer vision capabilities, including the ability t

This project is a toolkit and API designed for parsing, manipulating, and visualizing image annotations for computer vision tasks. It provides a programming interface to load and organize Common Objects in Context annotations, specifically for object detection, image segmentation, and keypoint estimation. The library includes tools for converting formatted JSON files into data structures that support the analysis of pixel-level masks and skeletal markers. It enables the visual verification of ground truth accuracy by rendering bounding boxes, segmentation masks, and keypoint markers directly

X-AnyLabeling is an AI-assisted annotation platform and computer vision labeling tool. It provides an interface for annotating images and videos using polygons and rectangles to create training sets for machine learning models. The project distinguishes itself through the integration of external AI models via a plugin-based inference backend, allowing for automated generation of candidate labels and the execution of specialized tasks like pose estimation and object detection. It also functions as an optical character recognition tool for extracting text and layout information from document im

CVAT is an open-source, web-based platform designed for annotating images, videos, and 3D point clouds to create high-quality training datasets for machine learning. It functions as a containerized server that orchestrates the entire lifecycle of computer vision data, from initial task creation and manual labeling to quality assurance and final dataset export. The platform distinguishes itself through deep integration with machine learning models, allowing users to deploy custom AI models as serverless functions for automated object detection, tracking, and skeleton annotation. It supports co

Frigate is a self-hosted network video recorder that functions as a private, local AI-powered vision engine. It manages video streams by performing real-time object detection, tracking, and classification directly on local hardware, ensuring that security monitoring and activity recording remain independent of cloud services. The system distinguishes itself through a modular, hardware-accelerated video pipeline that offloads intensive decoding and machine learning inference to dedicated GPUs, NPUs, or specialized accelerators like Coral TPUs and Hailo modules. It utilizes state-based object t

OpenCV is a comprehensive computer vision library designed for real-time performance and cross-platform deployment. It provides a native execution environment that leverages multi-threaded operations and automated memory management to handle intensive computational tasks, including image processing and machine learning model inference. The library distinguishes itself through a data-oriented matrix framework that utilizes proxy-based array abstractions to provide a consistent interface for multidimensional data. By employing factory-pattern algorithm interfaces and runtime type dispatching, i

mmocr is a PyTorch-based optical character recognition framework designed for training and deploying text detection, recognition, and key information extraction models. It serves as a comprehensive toolbox for scene text detection and recognition, providing specialized libraries for locating text regions and converting visual text into machine-encoded strings. The project distinguishes itself through a research framework for key information extraction and advanced text spotting capabilities. These include point-based spotting using transformers and the use of parameterized Bezier curves to id

ccv is a computer vision library written in C designed for high-performance visual analysis. It serves as a framework for image classification, object detection, and the identification of faces, pedestrians, and vehicles. The library distinguishes itself through hardware-accelerated vision and deep learning inference optimizations. It utilizes a quantized tensor processor to transform floating-point data into eight-bit integers and implements integer-quantized attention mechanisms to reduce memory bandwidth and increase data throughput. The project covers a broad range of capabilities, inclu

Libvips is a C-based image processing library designed to manipulate large visual assets through a low-memory, parallel processing pipeline. It functions as a streaming image processor that avoids loading entire files into system memory, enabling the handling of massive images in resource-constrained environments. The library distinguishes itself through a demand-driven architecture that constructs a deferred execution plan, computing only the necessary pixels for a final output. By utilizing a cache-friendly tiled processing model and memory-mapped file access, it minimizes latency and redun

This project is a collection of educational resources and implementation frameworks providing deep learning model recipes, code samples, and step-by-step guides for computer vision tasks. It organizes complex workflows into modular recipes and implementation guides to facilitate the building of image and video analysis models. The framework focuses on specialized vision capabilities, including an image similarity framework for fast retrieval and re-ranking, human pose estimation, and video action recognition. It also provides specific tools for crowd density estimation and document image clea

This is a real-time object detection framework built on the YOLOv3 architecture, implemented in PyTorch. It provides a complete pipeline for identifying and localizing objects in images and video using a single neural network pass, combining a Darknet-53 backbone with multi-scale feature pyramids and anchor-based bounding box prediction. The framework extends beyond basic detection to include instance segmentation, human pose estimation, and multi-object tracking across video frames. It offers a model export toolkit that converts trained models through ONNX to CoreML, TensorFlow Lite, and Ten

scikit-image is a Python image processing library and scientific image analysis toolkit. It provides a framework for digital image processing and computer vision, utilizing numerical arrays for pixel-level manipulations. The library enables the quantification of image properties and the detection of visual features, such as edges and blobs. It includes tools for image segmentation and the extraction of textures and patterns to characterize objects within visual data. Capabilities cover image manipulation through color space conversion, geometric transformations, and digital restoration. It a

This project is a modular PyTorch framework for training and evaluating object detection and instance segmentation models. It serves as a computer vision research tool and a deep learning inference engine designed to identify object locations, classes, and pixel-level masks within images. The framework implements a two-stage inference pipeline that utilizes region proposal networks and a symmetric mask-head architecture. It provides specialized capabilities for instance segmentation, object bounding box detection, and human pose estimation via anatomical keypoint detection. The system includ

This project is a comprehensive computer vision library for the PyTorch ecosystem, providing a standardized collection of neural network architectures, datasets, and high-performance transformation utilities. It serves as a foundational framework for building, training, and deploying deep learning models, offering a centralized model registry that allows developers to instantiate architectures with pre-trained weights for tasks such as image classification, object detection, and semantic segmentation. The library distinguishes itself through its modular approach to data and compute management

This project is a comprehensive toolkit designed for the full lifecycle management of large language and multimodal models. It functions as a unified orchestrator that handles the entire development process, ranging from dataset preparation and supervised fine-tuning to advanced reinforcement learning alignment and production-ready inference deployment. The platform distinguishes itself through a specialized reinforcement learning library that supports complex optimization algorithms, including group relative policy optimization and leave-one-out techniques, to improve model instruction-follo

This project is a modular research toolkit designed for developing, training, and evaluating deep learning models for object detection, segmentation, and video instance tracking. It provides a flexible training engine that manages complex neural network execution, including distributed training, custom lifecycle hooks, and weight optimization. The framework is built around a hierarchical configuration system that allows users to define architectures, data pipelines, and training hyperparameters through composable, inheritable files. The project distinguishes itself through its highly modular

LAVIS is a multimodal large language model framework and vision-language model library. It provides tools for training and evaluating models that integrate visual, textual, and audio data, serving as a cross-modal feature extractor and a zero-shot visual reasoning engine. The framework distinguishes itself by using frozen-backbone integration, where pretrained encoders remain non-trainable while lightweight adapter layers are updated. It employs cross-modal feature alignment to map different representations into a shared embedding space and utilizes a modular model wrapper to swap vision and

All-in-one training for vision models (YOLO, ViTs, RT-DETR, DINOv3): pretraining, fine-tuning, distillation.

Kornia is a differentiable computer vision library and cross-framework tensor vision toolset. It implements vision operations as differentiable tensors to enable integration into deep learning pipelines and supports the transpilation of operations across PyTorch, TensorFlow, JAX, and NumPy. The project provides specialized toolsets for geometric vision and stereo depth, including algorithms for 3D scene reconstruction, camera calibration, and pose estimation. It further distinguishes itself as a differentiable image augmentation framework, applying random geometric and color transformations w

Industry-strength Computer Vision workflows with Keras

CLIP is a neural network architecture designed to map visual and textual data into a shared latent vector space. By utilizing transformer-based feature extraction and multi-modal tokenization, the system aligns images and natural language strings, enabling cross-modal similarity analysis and semantic classification. The project functions as a zero-shot classification engine, identifying image content by calculating the cosine similarity between visual features and arbitrary text labels without requiring task-specific retraining. Beyond inference, it serves as a research toolkit for evaluating

MoviePy is a Python video editing library and automated video processor designed for programmatically cutting, concatenating, and manipulating video and audio files. It serves as a non-linear video editor and an interface for FFmpeg to handle the reading, writing, and conversion of diverse media formats and codecs. The library enables automated video composition through the layering of multiple video and audio streams using transparency and coordinate-based positioning. It supports dynamic content generation by inserting text overlays and performing custom video frame processing where raw fra