SBIR Phase I: Visitor Demographic Analysis in Shopping Malls using Computer Vision

小型企业创新研究(SBIR)第一阶段项目的更广泛影响/商业潜力将来自于构建一个计算机视觉平台，该平台将帮助使社区更有用，对其居住者更具吸引力，而不考虑人口统计数据，并以购物中心为试点应用。未来的购物中心将模糊社区中心和零售空间之间的界限，因为购物中心将成为人们不仅购物，而且还将生活、工作和娱乐的混合用途空间。为了实现这一转变，购物中心公司需要有关购物中心游客的详细人口统计信息。该公司的技术将改进最先进的行人运动和人口分类，而无需收集任何个人身份信息。这将使购物中心运营商能够了解和改善他们的租户结构，从而更好地在全国范围内培育社区。产生的数据还可以通过革命性的房地产分析而使其他行业受益：政府可以使用人口数据来分析不断变化的人口，城市规划者可以使用行人交通信息来使城市更清洁、更有效率。这项技术将使人们能够更详细地了解不断变化的社区以及如何帮助它们蓬勃发展，而不会像基于视频的技术那样通常会损害隐私。该公司打算在数据收集和管理方面保持隐私保护实践的领先地位。这个小型企业创新研究(SBIR)第一阶段项目将解决视频分析中的三个关键问题，以获取人口统计信息，同时避免收集个人身份信息，即：(1)公共环境中的行人人口统计分类，(2)准确的群体检测和路径(运动)分析，以及(3)使用网络摄像头图像进行稳定的路径分析。目前还没有大规模的系统能够准确地从网络摄像机图像中提取人口统计信息和路径信息。这项拟议的技术将分析遍布购物中心的摄像头网络的视频数据，提取人口统计的人流模式以生成人口统计和路径数据，然后永久删除视频。开发有效地从密集环境中出现的闭塞中恢复的技术对于该项目的商业采用至关重要。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project will result from building a computer vision platform that will help make communities more useful and appealing to their occupants regardless of demographic data, with shopping malls as the pilot application. Shopping malls of the future will blur the lines between community center and retail space as malls become mixed-use spaces where people will not only shop, but will also live, work, and play. To make this transition, shopping mall companies need detailed demographic information regarding mall visitors. The company’s technology will enable improved state-of-the-art pedestrian movement and demographic classification, without collecting any personally identifiable information. This will enable shopping mall operators to understand and improve their tenant mix, and thereby better foster communities nationwide. The data generated could also benefit other industries by revolutionizing property analysis: governments can use demographic data to analyze changing populations and urban planners can use foot-traffic information to make cities cleaner and more efficient. The technology will enable a more detailed understanding of changing communities and how to help them thrive, without the privacy compromises that usually accompany video-based technologies. The company intends to remain at the forefront of privacy-preserving practices with regard to data collection and management.This Small Business Innovation Research (SBIR) Phase I project will solve three crucial problems in video analytics for obtaining demographic information, while avoiding the need to collect personally identifiable information, namely: (1) classification of pedestrian demographics in a public setting, (2) accurate group detection and path (motion) analysis, and (3) stable path analysis using network camera imagery. There are currently no large-scale systems capable of accurately extracting both demographic and path information from network camera imagery. The proposed technology will analyze video data from a network of cameras placed throughout a shopping mall, extract demographic foot-traffic patterns to generate demographic and path data, and then permanently delete the video. Developing technology that effectively recovers from occlusions that arise in dense environments is crucial to the commercial adoption of the project.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.