Header image

Computer Science and Moves Institute
Naval Postgraduate School, MOVES, mailstop WA-265
700 Dyer Road Monterey, CA, 93943-5001
Tel: (831)-656-3443
 
 

Research

marines

Marine localization, pose estimation and classification. We consider the problem of detecting and classifying marines postures in still images, where the background is unconstrained, cluttered and not modeled. This imposes the detection of a large number of different objects and different views. We are focusing our efforts in a multi-class boosting procedure that favors discriminative features shared by multiple objects and views. As opposed to traditional methods, the approach implemented scales logarithmically with growing number of classes.Click here to see a video.

By Juan Wachs and Mathias Kolsch.

IR track  

Tracking objects at sub-pixel resolution. Robust video surveillance systems require detection of events at sub-pixel resolution to gain practicability. This work involves background modeling, using both spatial and intensity distributions charactersitcs. The first, assumes a normal model for background and noise intensity distribution, while the last is captured by an on-line estimate of the neighborhood structure affected by the sub-pixel event.

By Juan Wachs, Mathias Kolsch and Kevin Squire.
detect apples  

Enhancement of Sensing Technologies for Selective Tree Fruit Identification and Targeting in Robotic Harvesting Systems. A vision system that recognizes occluded apples within a tree canopy using IR and color is developed. The optimal registration parameters for affine transformation are obtained using maximization of mutual information. Haar features are then applied separately to color and IR images, and finally, a voting scheme which reduces false alarms without affecting the recognition rate is proposed.

By Juan Wachs, Victor Alchanatis, Helman Stern and Tom Burks.
surgeon using Gestix  

A Gesture-based Tool for Sterile Browsing of Radiology Images. The use of doctor-computer interaction devices in the operation room (OR) requires new modalities that support medical imaging manipulation while allowing doctors' hands to remain sterile, supporting their focus of attention, and providing fast response times. "Gestix," is a vision-based hand gesture capture and recognition system that interprets in real-time the user's gestures for navigation and manipulation of images in an electronic medical record (EMR) database. Click here for details

By Juan Wachs, Helman Stern, Yael Edan, Michael Gillam, Jon Handler, Mark Smith and Craig Feied.
 

Optimal Hand Gesture Vocabulary Design Methodology for Robot Control. A global approach to hand gesture vocabulary (GV) design is proposed which includes human as well as technical design factors. The human centered desires (intuitiveness, comfort) of multiple users are implicitly represented through indices obtained from ergonomic studies representing the psycho-physiological aspects of users. The main technical aspect considered is that of machine recognition of gestures. We believe this is the first conceptualization of the optimal hand gesture design problem in analytical form.

By Juan Wachs, Helman Stern and Yael Edan.
 

Tele-Gest Project: Hand Gestures Telerobotic Control. This project describes a telerobotic gesture-based user interface system using visual recognition. Experimental results showed that the system satisfies the requirements for a robust and user-friendly input device. The Fuzzy C-Means algorithm provided enough speed and sufficient reliability to perform the desired tasks. Although gestures were recognized quickly and sent in packet form through the internet, successful execution of the commands could not be verified until the image of the robot environment was received at the user interface.

By Juan Wachs, Uri Kartoun, Helman Stern and Yael Edan.