ATCam: Deep Neural Network-based Computer Vision Model of Detecting and Intervening Special Needs of Disabled People to Substantially Improve the Accessibility and Usability of Public Facilities

Abstract

ATCam (Assistive Technology Camera) is based on deep neural network computer vision technology to automatically detect and provide context-sensitive appropriate intervention to people with special needs in the context of using a variety of public facilities. The system comprises four modular components: sensing, identifying, decision-making, and feedback/communication modules. The identifying module uses deep neural network computer vision to detect individuals with disabilities based on assistive devices they use. The decision-making module recognizes intent and special needs to determine appropriate interventions. Laboratory trials demonstrated high accuracy and precision.