Correction of Optical Flow using Characteristic of Wrinkles on Palms

This paper describes a method of correcting the velocity vector by using the characteristics of the wrinkles in palms for hand gesture input for remote controllers. First, the HSV color system is used for hand images because this color system is not affected by changes in the lighting conditions. Next, an input method by switching the referenced parameters used for recognition when the range of acquisition of the images differs is proposed. Additionally, the block-matching method is applied to detect the motion vector because the ability to handle changes in brightness is required for consumer use environments. Moreover, how to reduce computational complexities and modify the detection accuracy of the rate vector are discussed. Finally, an experiment of operating a user interface based on the proposed method was conducted by using a practical system with an embedded MCU and a small camera. Consequently the effectiveness of the proposal was successfully confirmed