health

A Carnegie Mellon University (CMU) press release announced that Professor Takeo Kanade will lead an effort at CMU to study behavioral imaging for diagnosis of austism. In the research, computer vision will be used to analyze eye gaze, facial expression, and body posture to determine whether early signs of autism can be identified in young children. According to Kanade, "We aim at developing a new suite of technologies for imaging and understanding human behaviors, or behavior imaging. Just as medical imaging revolutionalized medicine and science for the body and its actions, behavior imaging will revolutionalize medicine and science for the mind and its activities."

Researchers at the University of California, Irvine, have developed a system to help young children with movement disabilities learn to control a motorized wheelchair. The system uses computer vision to follow a marked path on the floor, and guides the child's hand using a haptic force feedback joystick. The child tries to catch a small mobile robot that does a dance when it is caught. Early results for this method reported in the Journal of NeuroEngineering and Rehabilitation are promising. More information is also available in a Medical News Today article.

The Lung Cancer Alliance is running a project called Give A Scan, in which patients donate their CT scans and medical records to a database that is open to the public. The aim is to grow a dataset that researchers can use to develop new detection methods and treatments for lung cancer. The project has received support from Kitware which develops open-source software for image analysis and visualization.

Submitted by Boaz Super

The National Eye Institute and the National Institute of Aging are sponsoring LookTel, object recognition software to assist the blind and visually impaired. According to the LookTel website, the app will recognize packaged goods, money, CDs, DVDs, text, barcodes, landmarks, and medication bottles. The visually impaired person uses a Windows Mobile device; images are sent to a PC or laptop for recognition. A beta version is expected in Spring 2010.

Dr. Gang Luo of the Schepens Eye Research Institute at Harvard reports on several studies of computer vision enabled assisstive devices for visually impaired people. One device consists of see-through augmented reality glasses in which a minified version of the outlines of the scene are superimposed on the scene for patients with tunnel vision. Another device allows the option of two types of enhancements for patients with central vision loss: magnification and wideband enhancement. A third device assists patients with night blindness.

Imprivata, a company that provides secure login and authentication systems, has announced a new product that locks a workstation when a user walks away and automatically re-authenticates the user when he or she returns. The product, called OneSign Secure Walk-Away, uses computer vision to identify users and determine their presence.

Badri Roysam and Andrew Cohen at Rensselaer Polytechnic Institute (RPI) have developed a computer vision system that can predict with 99% accuracy whether stem cells will divide into self-renewing cells or terminal cells. The predictions, which are based on measuring cell movements, are made in real-time, so that the fate of the cells is known before they divide. The system can also predict characteristics of the specialized cells that result.

VTT, a contract research organization in Finland, has developed a fast method of image analysis to detect Alzheimer's disease in MR brain scans. Previous MR image assessment methods took from 15 minutes to several hours; the new method takes three minutes. The segmentation method, which uses expectation maximization (E-M) and graph cuts, automatically calculates the volume of the Hippocampus. More information is available in a press release and a Neuroimage article.

Seeing Machines Limited, maker of eye and face tracking systems, has released a new product. The TrueField Analyzer (TFA) measures pupillary response to diagnose glaucoma via a non-contact test.

SIMI Reality Motion Systems GmbH of Munich, Germany, has developed motion capture software for orthopedic diagnosis. The software analyzes video data to generate information about joint angles, accelerations, torque and stress. Doctors can then use the information to evaluate the patient’s musculoskeletal performance and offer a tailored treatment.