Keypoint Detection

Estimate human pose in images using pretrained HRNet keypoint detector or train custom object keypoint detector

Keypoint detection, also known as keypoint localization or landmark detection, is a computer vision task that involves identifying and localizing specific points of interest of an object in an image. In computer vision tasks, keypoints can represent human body joints, facial landmarks, or salient points on objects. Computer Vision Toolbox™ supports deep learning-based approach for keypoint detection in objects using high resolution deep learning network (HRNet). HRNet is based on convolutional neural networks (CNNs). For more information about the HRNet keypoint detector, see Getting Started with HRNet.

Keypoint detection provides essential information about the location, pose, and structure of objects or entities within an image, playing a critical role in computer vision applications such as:

  • Pose estimation

  • Object detection and tracking

  • Facial analysis

  • Augmented reality

Keypoint detection on a group of people

You can also train a custom object keypoint detector, or use transfer learning to modify a pretrained keypoint detector and fine-tune it for your application. For more information on transfer learning, see Deep Learning: Transfer Learning in 10 lines of MATLAB Code.

Apps

Image LabelerLabel images for computer vision applications
Video LabelerLabel video for computer vision applications

Functions

expand all

hrnetObjectKeypointDetectorCreate object keypoint detector using HRNet deep learning network (Since R2023b)
trainHRNetObjectKeypointDetectorTrain HRNet object keypoint detector (Since R2024a)
insertObjectKeypointsInsert object keypoints in image (Since R2023b)
loadHRNETObjectKeypointDetectorLoad HRNet object keypoint detector model for code generation (Since R2023b)

Topics

  • Getting Started with HRNet

    Learn high resolution network (HRNet) basics.

  • Deep Learning in MATLAB (Deep Learning Toolbox)

    Discover deep learning capabilities in MATLAB® using convolutional neural networks for classification and regression, including pretrained networks and transfer learning, and training on GPUs, CPUs, clusters, and clouds.

Featured Examples

Hand Pose Estimation Using HRNet Deep Learning

Hand Pose Estimation Using HRNet Deep Learning

Detect keypoints in a human hand and estimate the hand pose using the HRNet deep learning network.

Multi-Object Tracking and Human Pose Estimation

Multi-Object Tracking and Human Pose Estimation

Detect multiple people, track them, and estimate their body poses in a video by using pretrained deep learning networks and a global nearest-neighbor (GNN) assignment tracking approach.

Deep Learning-based Human Pose Estimation for Squat Analysis

Deep Learning-based Human Pose Estimation for Squat Analysis

Use pretrained deep learning networks to estimate human body pose and perform squat analysis.

Spacecraft Pose Estimation Using HRNet Keypoint Detector and PnP Solver

Spacecraft Pose Estimation Using HRNet Keypoint Detector and PnP Solver

Estimate keypoints and pose of spacecraft.