Predetermined voltage is applied to a liquid crystal lens by a liquid crystal lens driver. Image signals are generated based on an optical image passed through the liquid crystal lens during transient response operation caused by application of the predetermined voltage, and plural focus signals are extracted by sampling the image signals of the area from which the focus signals are extracted at predetermined cycles. Levels of the extracted focus signals are compared to determine the maximum value of the focus signal. Thus, with the liquid crystal lens, by making use of the transient response operation of the liquid crystal lens, a focus point can be detected in sufficient speed.