Computer Vision System Toolbox

Design and simulate computer vision and video processing systems

Computer Vision System Toolbox™ provides algorithms, functions, and apps for designing and simulating computer vision and video processing systems. You can perform feature detection, extraction, and matching; object detection and tracking; motion estimation; and video processing. For 3-D computer vision, the system toolbox supports camera calibration, stereo vision, 3-D reconstruction, and 3-D point cloud processing. With machine learning based frameworks, you can train object detection, object recognition, and image retrieval systems. Algorithms are available as MATLAB® functions, System objects, and Simulink® blocks.

For rapid prototyping and embedded system design, the system toolbox supports fixed-point arithmetic and C-code generation.

Feature Detection and Extraction

Image registration, interest point detection, extracting feature descriptors, and point feature matching

Object Detection and Recognition

Object detection, recognition, block matching, background estimation, bag of features

Object Tracking and Motion Estimation

Optical flow, activity recognition, motion estimation, and tracking

Camera Calibration

Camera calibration and stereo vision

Multiple View Geometry

3-D information extraction from 2-D images, stereo rectification, depth estimation, 3-D reconstruction, triangulation, and structure from motion

3-D Point Cloud Processing

Downsampling, denoising, transform, visualize, register, and fit geometrical shapes of 3-D point clouds

Analysis and Enhancements

Statistics, FIR filtering, frequency and Hough transforms, morphology, contrast enhancement, and noise removal

Input, Output, and Graphics

Importing, exporting, color space formatting, conversions, display, annotation

Code Generation and Third-Party Support

C Code generation, OCR language data support, OpenCV interface, fixed-point data type support, and System objects