A liquid crystal lens driver is used to apply a predetermined voltage to a liquid crystal lens. An image signal is generated based on an optical image having passed through the liquid crystal lens operating in transient response because of the application of the predetermined voltage. The generated image signal is sampled at predetermined intervals to extract a plurality of autofocus signals. A comparison is made among the levels of the extracted autofocus signals to determine the maximum value of the autofocus signals. In this way, the liquid crystal lens and the transient response operation thereof are used to detect a focus point with sufficient rapidity.