Recorded video segments are retrieved and displayed through a user interface that employs gestures. The user provides gestures by moving a mouse, and the resulting signals indicate a diagram. Data identifying the diagram is used to access a data structure to obtain data indicating which of a set of interactive line drawings includes features most closely resembling the diagram. The interactive line drawing with features closest to the diagram can then be displayed. When the user selects a part of the line drawing, a menu is displayed that includes a description of the video segments that relate to that part. To assist the user in providing gestural input, a preliminary diagram is displayed, reducing the amount of information the user must provide. The user can change parts of the diagram to obtain a modified diagram resembling features that appear in a desired line drawing. The diagrams and interactive line drawings can be displayed on a workstation display screen, while the video segments can be displayed on a video monitor screen by a video disk player under control of the workstation CPU.