A method for identifying one or more image frames (100) from a series (104) of video image frames, in order to print, store or flag the selected frames. The method calculates a gradient value for each image frame in a video shot (102) series, identifies one or more gradient value peaks (324) in a corresponding range of a raw data series (302), and selects the image frames (110) corresponding to the identified gradient value peaks. Two types of frames (202, 206) are identified as being suitable for printing. A first type (202) comprises frames with a high gradient content, where the image is sharply defined. A second type of frame (206) also has good focus, but is representative of periods of relative stillness, or lack of motion between consecutive image frames (100) during the video shot series (104).