Computer Vision

Computer Vision is the branch of computer science concerned with the analysis of images to extract information about the world. This is the same function that the human visual system provides (although perhaps accomplished through different mechanisms). As sensor and computer hardware drops in cost, these visual functions can become features in a wide range of products where they provide automatic, fast, convenient, and precise alternatives for tasks that were previously manual.

Much of the computer vision research at MERL is focused on the area of surveillance. For example, MERL has pioneered a state of the art approach to detecting object classes such as human faces in cluttered scenes. This approach uses a powerful machine learning framework to automatically build very fast object detectors given a set of positive and negative examples of the object class. The same approach has been successfully applied to the problems of pedestrian detection, facial feature finding, face recognition, and gender and race classification. Last fall this work was honored with the Marr award at a primary computer vision conference. Another focus in the surveillance area is object tracking in video. Some of the work in tracking has used stereo cameras to track objects in 3-D. Other work has looked at the problem of tracking objects across different cameras in multi-camera systems. Once having detected and tracked objects, we now strive to analyze the activity or event taking place.

The following project descriptions describe the many projects going on at MERL in the area of computer vision. They include work on biometric systems tracking systems, audio-visual event detection, intelligent video browsing systems and fusion of imaging sensors. These systems are being applied to many areas of Mitsubishi Electric's businesses such as surveillance and security, consumer products (cell phones and DVD players) and elevators.

Past Projects:
    3D Face Recognition
    3D from Video
    A Fast Algorithm for Depth Segmentation
    A Projector as a Novel Type of Motion Sensor
    Artificial Retina Skunkworks
    Audio-Visual Event Detection for Consumer and Surveillance Video
    Biometrics Using Stereo Vision
    Body Tracking from Single-Camera Video
    Building 3D Models of the Human Head
    Camera Network Calibration
    Component-Based Face Recognition
    Computer Human Observation (CHO)
    Computer Vision for Computer Games
    Context-Aware Pan-Tilt-Zoom Cameras
    Correctness of Belief Propagation in Bayesian Networks with Loops
    Covariance Tracking
    Data-mining and Recommending
    Detecting Visual Tags
    Diamond3D Computer Vision Library
    DiamondBuild
    DiamondClassify
    Digital Merchandising
    Dimensionality Reduction
    Easy Calibration of a Projector
    Ensemble Tracking
    Event Detection
    Exploiting the generic viewpoint assumption to estimate scene parameters
    Face Based Browsing for Surveillance Applications
    Face Detection/Gender & Race Classification
    Factorized Local Appearance Models
    Fast super-resolution method
    Generalized Belief Propagation Algorithms
    Hand-Held 3D Scanning Using Computer Vision
    Hand-Held Projectors for Augmented Reality
    Happy and Sad Face Classifiers
    Heli-Tele
    Human Activity Determination
    Hypercuts: Boosted Dyadic Kernel Discriminants
    Image Retrieval with Multiple Regions-of-Interest
    Interactive Surroundings
    Iris Recognition from 1-2 Meters
    Learning Concise Models of Visual Activity
    Learning Normal Activity and Detecting Anomalies
    Learning low-level vision
    Low Cost Projector Mosaic
    Low-Frame-Rate Tracking
    MERL Optic Touch
    Manifold of Faces
    Mitsubishi Electric's Intelligent CMOS Image Sensor (ICIS)
    Motion-Based Optical Sensing with Multiple AR Cameras
    Moving Cast Shadow Detection
    Multi-Camera Systems
    Multi-Projector Imagery on Curved Screens
    Multilinear Face Models
    Object Tracking & Understanding
    Observing and Classifying the Activity of a Vehicle Driver
    PEP: Performance Evaluation Platform for Object Tracking Methods
    PalmCam - Digital Camera for PDA
    Pedestrian Detection
    Personal Digital Historian (PDH)
    Personal Eyewitness CarCam - Vehicle Accident Video Recorder
    Probabilistic Modeling for Face Recognition
    Projector
    Recovery of 3D Shape from Images
    Road Extraction for Satellite Imagery
    SCAR - Super Cheap Artificial Retina Evaluation Board
    Scene Analysis using Camera Arrays
    Shadow Puppetry
    Single-Axis Multi-Parameter (SAMP) Camera
    Spectral Bounds for Sparse PCA and Sparse LDA
    Stereo Computer Vision for Observing People
    Super-Resolution Using a Markov Network Approach
    Support Vector Learning for Gender Classification
    Surface Reconstruction
    Surveillance Architecture
    System Identification for Video Texture
    Tangible Intermediaries
    Television Set Controlled By Hand Gestures
    Unusual Event Detection
    UrbanMatch and AerialMatch - Image Matching Applications
    Video Object Segmentation
    Video Object Tracking
    Video Surveillance with NPR Image Fusion
    Video Warehousing and Face Classification Visualization
    VideoRule - Automatic Integration of Video in Databases
    Visual Tracking & Recognition with Particle Filters
    Visual Tracking of Flexible 3D Surfaces
    Visualization & Layout for Image Libraries
    Waviz Background Models
    Wheelchair Detection Using Stereo Vision
    iLamps: Intelligent, Locale-aware, Mobile Projectors