|
|
 Intelligent Traffic Intersection
Video-based target detection at signized intersections has many advantages over traditional inductive loops, such as easy set up, easy definition of detection zones, easy modification of detector settings, and lower maintenance cost. However, several common problems need to be resolved, including low lighting conditions due to adverse weather conditions.
The main goal of this project is to design the video-based target detection algorithm by fusing visible with infared (IR) data to improve both accuracy and robustness, especially under adverse conditions (i.e., night-time, snow, rain, fog, etc.).
We describe a traffic scene background model, a background subtraction process and on-line background update process that we have developed based on the support vector regression (SVR). Following an incremental support vector regression, the developed background model is on-line updated whenever a training input of intensity values is added to the training set. Such scheme enables the developed background model to be adaptive to the intensity variance of images caused by the changes of outdoor environment (i.e. illumination changes for visible image and variance of the thermodynamic and atmospheric conditions for infrared image.)
|
| |
|
 Automatic Background Modeling for Visual Surveillance
The final goal for many visual surveillance systems is automatic understanding of events in a site. Higher level processing on video data requires certain lower level vision tasks to be performed. One of these tasks is the segmentation of video data into regions that correspond to objects in the scene. Issues such as automation, noise robustness, adaptation, and accuracy of the model must be addressed. Current background modeling techniques use heuristics to build a representation of the background, while it would be desirable to obtain the background model automatically. In order to increase the accuracy of modeling it needs to adapt to different parts of the same scene and finally the model has to be robust to noise.
|
| |
|
 Vehicle Detection Robust
and reliable vehicle detection is one of the most important
issues in any in-vehicle optical system, with applications
to driver assistance systems or autonomous, self-guided
vehicles. Several factors make on-road vehicle detection
very challenging including variability in scale, location,
orientation, and pose. The focus of this project, which is
funded by Ford Motor Company, is to develop a reliable
real-time on-road vehicle detection system.
|
| |
|
 Automatic Target Recognition Using Algebraic Functions of Views The main goal of this project is to improve the performance of Automatic Target Recognition (ATR) by developing a more powerful ATR frame work
which can handle changes in the appearance of a target more efficiently and robustly. The new framework will be built around a hybrid model of
appearance by integrating (1) Algebraic Functions of Views (AFoVs), a powerful mathematical model of geometric appearance, with (2) eigenspace
representations, a well known empirical model of appearance which has demonstrated significant capabilities in recognizing complex objects under no
occlusion.
This project is sponsored by The office of Naval Research (ONR).
|
|
 Surveillance In this project, we have been developing a system that acquires and processes images through one or multiple Camera Units monitoring certain area(s) via a Local Area Network (LAN) and is capable of combining information from multiple Camera Units to obtain a consensus decision. The system can be trained to detect certain type of intrusions, for example pedestrians, a group of pedestrians, vehicles, pets etc., and minimizes false alerts due to other non-interested intrusions. As a case study, we are using this system to detect Pedestrian/Vehicle in an observation area.
|
| |
 Computer Vision Technologies for Effective Human-Computer
Interaction in Virtual Environments
Developing efficient and effective virtual environment technologies to provide sophisticated training sessions for astronauts and to help ensure safety is a high priority to NASA. Our main goal in this project is to Advance NASA's virtual environment technologies by developing and demonstrating computer vision non-contact techniques for efficient and effective human-computer interaction in virtual environments. This project is funded by NASA and is a joint effort with BioVIS Lab at NASA Ames Research Center.
|
| |
 Voting Based Computational Framework for Motion Analysis This work addresses the problem of visual motion analysis, by formulating
it as an inference of motion layers from a noisy and possibly sparse point
set in a 4-D space. The approach is based on a layered 4-D representation
of data and a voting scheme for token communication, within a tensor
voting computational framework. In the 4-D space of image positions and
velocities, moving regions are conceptually represented as smooth surface
layers, and are extracted through a voting process that enforces the
smoothness constraint while preserving motion discontinuities. This
non-iterative framework determines an accurate scene description in terms
of motion regions, boundaries and dense image velocities, without using
any parametric motion model - the only criterion used is the smoothness of
image motion.
|
| |
|
 Face Detection in the Infrared Spectrum
Face detection is an important prerequisite step for successful face recognition. The performance of previous face
detection methods reported in the literature is far from perfect and deteriorates ungracefully where lighting conditions
cannot be controlled. We propose a method that outperforms state-of-the-art face detection methods in environments
with stable or variable lighting conditions. The approach capitalizes upon the near-IR phenomenology for skin detection.
|
| |
|
 Gender Classification
Computer vision systems for surveillance and human computer
interaction must be able to process human faces in a variety
of ways. Gender information can be used to enhance existing Human Computer Interaction (HCI) systems
but also can serve as a basis for passive surveillance and control
in "smart buildings" (e.g., restricting access to certain areas
based on gender) and collecting valuable demographics (e.g., the
number of women entering a retail store on a given day). The goal of this project is to develop a robust gender classification system. The key idea of our approach is using subset feature selection.
|
|
