Flow : a framework for reality-based interfaces

[摘要] This thesis outlines the design, development and evaluation of a computer vision framework known as Flow. Flow utilizes a wearable head mounted camera to observe user interactions, providing recognition of objects, locations, people, and gestures via machine vision and understanding. This supports the creation of a hybrid interaction environment, allowing users to access digital functionality and actions based on their interactions within the physical environment. Flow highlights the potential of cross device, environment interactions by providing the architecture to connect between existing platforms and devices. Example applications scenarios are provided to evaluate the framework. Each demonstrates the versatility of the system in supporting human computer interaction and prototyping of ubiquitous systems. Insight through user feedback and a technical evaluation is also provided. The work outlined in this thesis demonstrates the potential of passive observational computer vision systems in the augmentation of user interactions and the potential for further dissolved computing environments.

[发布机构] Massachusetts Institute of Technology