The Algorithms Research And Micro-system Implementation On Depression Trend Detection

Depression will become the most powerful killer in the future. With the development of the technologies, the researches of brain cognitive function are also developing. The known conclusion prompted us that the depression cause is brain's activity decrease on temporal lobe and prefrontal region. One of the most effective approaches of controlling depression is that to detect the depression according to brain mechanism, and to concern depression in real time.The target of this thesis is to detect the brain mechanism indirectly, and to research the typical characteristic (smile vs calmness) of infrared auricular signal by our analysis system (the personal computer system and the advanced RISC machine micro-system). Based on the acquisition of auricular signal using infrared sensor, I chose the difference between smile and calmness from the participants to detect the depression trend. We get the energy in milli-hertz with the improved algorithm of wavelet transform, and we analyzed detection signals'character of nonlinear with the algorithm of multi-fractal. There are 20 volunteers attaching to our experiment to help us validating the feasibility of our methods. At last, we implemented the depression detection in ARM micro-system. In this system, we used touch-screen to control our program and to display the result, and we embedded the algorithms into interrupted serve program.The finished experiments results had proved that: (1) It had been proved to be successful that using the deference between calmness and smile to detect the depression trend. (2) The two features of wavelet and wavelet plus multi-fractal are all suitable for detecting depression on our system. (3) Using the wavelet feature follows the brain cognitive law that is the calmness energy of human with good health being higher than their smile energy. (4) The feature of wavelet plus multi-fractal can be used to analyze the limited discrete time signal. The analysis about data of calmness and smile suggest that the calmness average value of dimension is lower than smile for the man with depression. (5) The depression detection validity touched up to 80 percent for ARM micro-system platform or personal computer system.The designing of our experiment will become the important base of depression detection apparatus in the field of SoC.