Spatial sketch to compose internet images with RealSense 3D camera

Digital content creation become an indispensable part of our lives. In addition, massive internet images provide us abundant creation material. To construct a computer-aided graphics creation system, interaction and convenience are the critical factors. With somatosensory interactive technology, we can interact with our devices in a more natural and intuitive way, and gain immersive experience at the same time. In this paper, we present a sketch driven image creation system by retrieving and composing internet images automatically. These images are retrieved according simple freehand sketch annotated with text labels. With the help of RealSense depth camera, user only need to make various hand gestures to draw the sketch in spatial and tag with the text labels by speech. Then our system will complete a series of tasks to generate image automatically, thus it greatly reduce the burden of painting and increase the fun and interactive friendliness.