3 dépôts
Methods like generalized focal loss to improve localization quality during training.
Distinct from Detection Model Validation: Distinct from model validation: focuses on training-time refinement logic rather than post-training evaluation.
Explore 3 awesome GitHub repositories matching artificial intelligence & ml · Bounding Box Refinement Techniques. Refine with filters or upvote what's useful.
PaddleDetection is an object detection framework designed for the end-to-end development, training, and deployment of computer vision models. It provides a comprehensive library of modular neural network architectures and pipelines that support object detection, instance segmentation, and multi-object tracking tasks. The project distinguishes itself through a configuration-driven approach that decouples model components like backbones and heads, allowing for the flexible assembly of custom vision workflows. It incorporates advanced techniques such as anchor-free detection logic, joint detecti
Refines bounding box quality during training using generalized focal loss techniques.
tensorrtx is a computer vision inference engine and model implementation library designed for graphics processor acceleration. It provides a framework for optimizing deep learning models through a GPU inference optimizer, a deep learning model converter for transforming weights from frameworks like TensorFlow and PyTorch, and a custom plugin library to implement operations not natively supported by the TensorRT API. The project distinguishes itself through a comprehensive collection of pre-defined network implementations, ranging from various YOLO versions and DETR transformers for object det
Provides post-processing logic to refine bounding box predictions and handle multi-output tensors after inference.
CenterNet est un framework de détection d'objets par points centraux et un pipeline de vision par ordinateur en temps réel. Il identifie les objets et les poses en prédisant les points centraux au lieu d'utiliser des boîtes d'ancrage (anchor boxes). Le système fonctionne comme un estimateur de boîtes englobantes 3D, un modèle d'estimation de pose humaine et un outil de détection d'objets en temps réel. Il traite le placement des articulations et les emplacements des objets comme des problèmes de détection de points centraux pour localiser les entités dans les images et l'espace tridimensionnel. Les capacités couvrent la détection d'objets 3D, l'estimation de points clés humains et l'analyse vidéo en direct. Le pipeline utilise un processus d'inférence feedforward à étape unique pour effectuer une analyse continue sur des webcams ou des fichiers vidéo.
Predicts local offsets to correct quantization errors and refine bounding box precision.