Shooting Assistance by Recognizing User's Camera Manipulation for Intelligible Video Production

Abstract

We propose a system which achieves cooperative capture assistance by camera manipulation recognition. Based on an experimental result on inexperienced users, the incremental interaction model, which the system and a user cooperatively shoot, was formerly proposed. The system based on the model compensates for user's lack of cinematographic knowledge or skills by relating affective information such as atmosphere or mood to capture techniques. When a user captures a shot after selecting a specific atmosphere, the system analyzes the current shooting image and the camera operation including the camera angle and the zooming speed. Then it gives guidance for better capture according to the analysis. The proposed system based on the model achieves an incremental interaction between the user and the system, evolving from user's unidirectional manipulation of the system. The system assists the user in reflecting user intention of the scene appropriately, therefore it enables the user to capture scenes more appropriately and effectively without specific cinematographic knowledge or skills. As a result, the user can acquire basic shooting skills smoothly and shoot more effectively.