Vision-based Resource Constrained Event Detection for Medical Smart Homes

Abstract

<p>As the number of elderly persons as well as their fraction of the total population<br />continues to rise, especially in the developed countries, providing an appropriate living<br />environment for them using smart home technology is rapidly gaining attention. Two<br />important tasks of a smart home technology are monitoring the daily activities and<br />the vital signs of the elderly to improve their quality of life and to monitor existing<br />or the onset of health abnormalities. In this thesis, we focus on the monitoring of<br />taking medicine by the elderly person using vision sensors (low-cost cameras). This<br />task is important since it helps both the person and the doctor in the treatment<br />of illnesses of elderly persons. The allocated resources of communication bandwidth<br />between the sensor nodes and the computational power, used for this task, affect<br />the implementation cost. Therefore, it is desired to develop an effective scheme<br />which efficiently allocates bandwidth and computational resources to achieve a high<br />reliability (detection performance) at low cost.<br />In this thesis, we have proposed two different approaches to solve this detection<br />and monitoring problem. As the input data are video frames, captured by cameras<br />from the same scene, the frames have inter-view redundancy. Taking advantage of<br />this inter-view redundancy, we proposed a video coding classification scheme based<br />on separate encoding and joint decoding, and have obtained significant compression<br />improvement compared to existing techniques. In the second approach, we studied<br />different parts of the detection and monitoring system to find an efficient design<br />for distribution of different event detection parts between the nodes and the central<br />processing unit so that the allocated resources are reduced. In this scheme, the<br />useful information of the frames are extracted in the form of their main features<br />such that decision making based on these features is the same as decision making<br />based on the raw frames. As a result, we could propose a new scheme which requires<br />significantly less bandwidth and computational resources while achieving the same<br />detection performance.</p>