An Embedded Lens Controller for Passive Auto-Focusing Camera Device Based on SOM Neural Network

Recently, camera has been widely used in the mobile device in order to provide the function of photograph. In camera's control techniques, passive auto-focusing (PAF) has become a tendency to improve the quality of image and the production cost. Basically, the sharpness measurement algorithm is used to measure the focused values of the scene captured. However, the conventional control methods are not very suit for PAF technique due to the full search is still used in PAF control, and it can not efficiently capture the image in a dynamic situation. To decrease the searching time needed, this paper presents a new controller based on SOM neural network. The proposed method is easily to be implemented as the embedded system or be designed as a slimmer and smaller mobile camera device by using a specific integrated circuit. In our study, the implementation is demonstrated by an integrated system which includes CMOS image sensor with adjustable lens, micro-controller and FPGA