AIR Option 1: Technology Translation: Automated Targeted Destination Recognition for the Blind with Motion Deblurring

这个PFI：AIR技术翻译项目专注于翻译计算机视觉技术的研究进展，以填补盲人辅助技术研究与实际需求之间的差距。该项目的目标是探索解决方案，从基础研究到生产便携式概念验证原型，帮助盲人用户从复杂环境中识别目标目的地，从而为盲人设计可穿戴的基于相机的定向辅助工具。该项目通过1)开发稳健的运动去模糊方法来处理戴着摄像头的盲人用户的不规则运动，以及2)根据盲人用户对路线和导航的请求来识别目标目的地，从而实现了这一目标。概念验证原型系统将开发通过太阳镜/帽子上的可穿戴摄像头和便携式计算机(迷你笔记本电脑或手机)捕获的视觉信息进行数据分析，而语音/声音输出将通过蓝牙耳机提供给盲人用户。新开发的算法和概念验证原型的性能将由盲人受试者进行评估。这一合作伙伴关系涉及不同领域的专业知识，包括大学研究、工业真实系统设计和开发，以及盲人和视障组织的盲人用户研究，以解决为盲人用户设计可穿戴式相机定向导航辅助设备中最具挑战性的问题(运动模糊和基于查询的目标识别)。专业知识的跨学科领域涉及将基于计算机视觉的辅助技术转化为可能导致竞争性商业现实的途径的潜力。这将带来新的革命性设计理念，即高性价比和便携式的基于相机的辅助设备，帮助视障人士实现与视力正常的人相当的功能移动性。预计未来10年潜在的经济影响将达到1.2亿美元，这将有助于美国在开发盲目寻路和导航技术方面的竞争力。这项研究对造福视障人士、促进他们融入和融入社会、提高他们未来的就业机会、在工作场所取得成功、独立生活以及经济和社会自给自足具有直接影响。

This PFI: AIR Technology Translation project focuses on translating the research advances in computer vision technologies to fill the gap between research and the real needs in assistive technology for blind people. The goal of this project is to explore solutions for the challenges of designing wearable camera-based wayfinding assistants for blind persons from fundamental research to produce a portable proof-of-concept prototype to help blind users recognize targeted destinations from complex environments. The project accomplishes this goal by 1) developing robust motion deblurring methods to handle the irregular motions of blind users who are wearing the camera and 2) recognizing targeted destinations based on the requests of blind users for wayfinding and navigation. The proof-of-concept prototype system will be developed with visual information captured via a wearable camera on sunglasses/hat and a portable computer (mini laptop or cell phone) for data analysis, while the speech/sound outputs will be provided to a blind user via a Bluetooth earpiece. The performance of the newly developed algorithms and the proof-of-concept prototype will be evaluated by blind subjects.The partnership engages different areas of expertise including university research, industry real system design and development, and blind user study from organizations for the blind and visually handicapped to provide solutions for the most challenging problems (motion blur and query-based targeted destination recognition) in the design of wearable camera-based wayfinding and navigation aids for blind users. The interdisciplinary areas of expertise pertain to the potential to translate the computer vision based assistive technology along a path that may result in a competitive commercial reality. This will lead to new revolutionary design concepts of cost-effective and portable camera-based assistive devices to help visually impaired people in achieving functional mobility comparable to people with normal vision. The potential economic impact is expected to be $120 million in the next 10 years, which will contribute to the U.S. competitiveness in developing technology for blind wayfinding and navigation. This research has a direct impact to benefit the visually impaired, improve their inclusion and integration into society, and enhance their future employment opportunities, success in the workplace, independent living, and economic and social self-sufficiency.