Object detection on multiple datasets with an automatically learned unified label space.

A package to read and convert object detection datasets (COCO, YOLO, PascalVOC, LabelMe, CVAT, OpenImage, ...) and evaluate them with COCO and PascalVOC metrics.

PiGallery: AI-powered Self-hosted Secure Multi-user Image Gallery and Detailed Image analysis using Machine Learning, EXIF Parsing and Geo Tagging

Scaling Object Detection by Transferring Classification Weights

YOLOv4 (v3/v2) - Windows and Linux version of Darknet Neural Networks for object detection (Tensor Cores are used)

Convert OpenImages labels to be used for YOLOv3

Open Images is a dataset of ~9M images annotated with image-level labels, object bounding boxes, object segmentation masks, visual relationships, and localized narratives.

Animals object detection such as deer, horse, and rabbit in diverse settings using YOLOv5

Train a custom semantic segmentation model using the Open Images Dataset and IceVision, an object detection framework that runs on top of PyTorch, FastAI, and PyTorch Lighting.

Training a model for Instance segmentation and object detection with MaskRCNN with TensorFlow on a custom selected dataset from the open image.

YOLO v3 on OpenImages dataset in CoreML with Vision implemented for iPhone iOS in Swift

Convert openimages v4 dataset to darknet train datas.

Tools to set up and download the dataset + Sample dataset

SugarKubes Public Documentation

Inceptionv3 model trained on OpenImages dataset, with code in Tensorflow

Google OpenImages V7 is an open source dataset of 9.2 million images annotated with image-level labels, object bounding boxes, object segmentation masks, and visual relationships.

Mahajan et al. and Yalniz et al. demonstrate the benefit of pre-training on large unlabeled image datasets using user tags as labels for weak-supervision. We take this approach further, considering whether the use of user tags in large labeled image datasets is beneficial.

This repo includes the implemetation of some of the state of the art object detectors on subsets of some of the most popular public datasets for object detection task.

Download sub part of open image dataset