Keypoint Detection

The COCO (Common Objects in Context) dataset is a large-scale object detection, segmentation, and captioning dataset. It is designed to encourage research on a wide variety of object categories and is commonly used for benchmarking computer vision models. It is an essential dataset for researchers and developers working on object detection, segmentation, and pose estimation tasks.

The MPII Human Pose Dataset for single person pose estimation is composed of about 25K images of which 15K are training samples, 3K are validation samples and 7K are testing samples (which labels are withheld by the authors). The images are taken from YouTube videos covering 410 different human activities and the poses are manually annotated with up to 16 body joints. Source: 2D/3D Pose Estimation and Action Recognition using Multitask Deep Learning Image Source: http://human-pose.mpi-inf.mpg.de/

The Pascal3D+ multi-view dataset consists of images in the wild, i.e., images of object categories exhibiting high variability, captured under uncontrolled settings, in cluttered scenes and under many different poses. Pascal3D+ contains 12 categories of rigid objects selected from the PASCAL VOC 2012 dataset. These objects are annotated with pose information (azimuth, elevation and distance to camera). Pascal3D+ also adds pose annotated images of these 12 categories from the ImageNet dataset. Source: Convolutional Models for Joint Object Categorization and Pose Estimation Image Source: Beyond PASCAL: A benchmark for 3D object detection in the wild

This dataset focuses on heavily occluded human with comprehensive annotations including bounding-box, humans pose and instance mask. This dataset contains 13,360 elaborately annotated human instances within 5081 images. With average 0.573 MaxIoU of each person, OCHuman is the most complex and challenging dataset related to human. Source: https://github.com/liruilong940607/OCHumanApi Image Source: https://github.com/liruilong940607/OCHumanApi

KeypointNet is a large-scale and diverse 3D keypoint dataset that contains 83,231 keypoints and 8,329 3D models from 16 object categories, by leveraging numerous human annotations, based on ShapeNet models. Source: https://github.com/qq456cvb/KeypointNet Image Source: https://github.com/qq456cvb/KeypointNet

ApolloCar3DT is a dataset that contains 5,277 driving images and over 60K car instances, where each car is fitted with an industry-grade 3D CAD model with absolute model size and semantically labelled keypoints. This dataset is above 20 times larger than PASCAL3D+ and KITTI, the current state-of-the-art. Source: ApolloCar3D: A Large 3D Car Instance Understanding Benchmark for Autonomous Driving Image Source: http://apolloscape.auto/car_instance.html

The General Robust Image Task (GRIT) Benchmark is an evaluation-only benchmark for evaluating the performance and robustness of vision systems across multiple image prediction tasks, concepts, and data sources. GRIT hopes to encourage our research community to pursue the following research directions: 1. General purpose vision models - GRIT facilitates the evaluation of unified and general-purpose vision models that demonstrate a wide range of skills across a diverse set of concepts. 2. Robust specialized models - GRIT simplifies and unifies quantification of misinformation, calibration, and generalization under distribution shifts due to novel concepts, novel data sources or image distortions for 7 standard vision and vision-language tasks. 3. Efficient learning - GRIT includes a restricted and an unrestricted track. The restricted track constrains the allowed training data to a selected but rich set of data sources that allows more scientific and meaningful comparison between models. This is meant to encourage resource constrained researchers to participate in the GRIT challenge and to spur interest in efficient learning methods as opposed to the dominant paradigm of training larger models on ever increasing amounts of training data. The unrestricted track allows much more flexibility in training data selection to test the capability of vision models trained with massive data and compute.

AwA Pose is a large scale animal keypoint dataset with ground truth annotations for keypoint detection of quadruped animals from images.

The researchers collected 3,500 images of Tilapia fish, with each image containing three fish in a small bowl. These images were manually annotated using Roboflow, a tool for creating and managing annotated datasets. Four keypoints were labeled on each fish: mouth, peduncle, belly, and back. While the primary goal was to measure fish length using the mouth and peduncle points, the additional keypoints (belly and back) were included to support potential future research, such as using girth to determine fish weight. This dataset was used to train a YOLOv8 model for keypoint detection, achieving high accuracy in identifying these crucial points on the Tilapia fish.

A multimodal dataset of radio galaxies and their corresponding infrared hosts.

Automating the creation of catalogues for radio galaxies in next-generation deep surveys necessitates the identification of components within extended sources and their respective infrared hosts. We present RadioGalaxyNET, a multimodal dataset, tailored for machine learning tasks to streamline the automated detection and localization of multi-component extended radio galaxies and their associated infrared hosts. The dataset encompasses 4,155 instances of galaxies across 2,800 images, incorporating both radio and infrared channels. Each instance furnishes details about the extended radio galaxy class, a bounding box covering all components, a pixel-level segmentation mask, and the keypoint position of the corresponding infrared host galaxy. RadioGalaxyNET is the first dataset to include images from the highly sensitive Australian Square Kilometre Array Pathfinder (ASKAP) radio telescope, corresponding infrared images, and instance-level annotations for galaxy detection.

TAMPAR is a real-world dataset of parcel photos for tampering detection with annotations in COCO format. For details see the paper and for visual samples the project page. Features are: - >900 annotated real-world images with >2,700 visible parcel side surfaces - 6 different tampering types - 6 different distortion strengths

The ViCoS Towel Dataset is a state-of-the-art benchmark for grasp point localization on cloth objects, specifically towels. Designed to advance research in robotic grasping and perception for textile objects, this dataset includes a collection of 8,000 high-resolution RGB-D images (1920×1080) captured with a Kinect V2 under a variety of conditions. Each image provides detailed depth information, making it ideal for training deep learning models and conducting thorough benchmarking. Object Diversity The dataset features 10 types of towels from the Household Cloth Objects. including various sizes and patterns such as big towels, checkered rags, and waffle rags. Each towel's corners are treated as potential grasp points for detection and localization. Object Positions Towels were positioned in 10 different configurations on a tabletop to simulate a range of visibility scenarios, from fully spread-out to crumpled and folded, providing diverse grasping conditions. Background Variability To enhance background variability, we used five different tabletop cloths or objects, including festive tablecloths and patterned tablecloths. This variation helps test algorithms under different background conditions. Lighting Conditions Images were captured under eight distinct lighting setups, ranging from no light to fully illuminated scenes with various shadow intensities. This variety helps in evaluating the robustness of grasping models under different lighting scenarios. Clutter Both cluttered and uncluttered scenes were included to create challenging conditions with occlusions. Various clutter items were added on the desk or directly on the towels to simulate realistic scenarios. ### Annotations Each visible corner of the towels in the images has been manually annotated as a potential grasping point. Annotations include a point label and the angle-of-approach, providing a total of 20,784 annotated points with corresponding angles. ### Training and Testing Splits The dataset is divided into training and testing subsets. Two towels (checkered rag small and cotton napkin) and one background (festive tablecloth) were reserved for testing, ensuring that these elements are not present in the training set. The dataset includes 5,120 training images and 2,880 testing images, covering all object configurations, lighting conditions, and clutter scenarios. ### Synthetic Training Data In addition to the real images, the dataset includes 12,000 synthetic images generated using the MuJoCo simulation environment. These synthetic images depict towels in varied positions and conditions, with different textures and lighting settings, providing further diversity for training purposes. For more details and access to the dataset, visit GitHub repository. CeDiRNet-3DoF: Center Direction Network for Grasping Point Localization on Cloths!