Taking self-portraits with a digital camera is a popular way to present oneself through photography. Traditional techniques for taking self-portraits, such as use of selftimers or face detection, provide only a modest degree of interaction between the user and camera. In this paper, we present an interaction technique that make novel use of image-processing algorithm to recognize hand motion gestures and provides user a natural way to interact with camera for taking self-portraits. User can perform nature gestures to control essential functions of camera and take self-portraits effectively. Three types of gesture (i.e., waving, eight-direction selection, and circling were identified and applied to develop a gesture user interface for controlling a Digital Single-Lens Reflex (DSLR) camera. Two experiments were conducted to evaluate the usability and performance of the gesture interface. The results confirmed that the usability of the gesture interface is superior to a self-timer and the proposed technique achieved about 80% accurate recognition of motion gestures.