The Design And Implementation Of Image Synthesis System Based On Hand-drawn Sketch

Image plays an important role in people’s daily life, production management, scientific research and other fields. With the rapid development of science and information technology, images sharing are more frequently and people’s requirements for images are more abundant. At the same time, an effective image shows not only more information to people than words, but also more intuitive and more convincing. But a single image has limitation in both space and time, so that it may not meet the people’s requirements. Therefore, the requirement for image synthesis is derived. Image synthesis is a method that by synthesizing multiple images in a scene to a new image to exhibit a scene description which a single image doesn’t have.At the present, image synthesis mostly requires the user to provide the original image. This is a high time cost. So we developed an image synthesis system based on hand-drawn sketches. The system uses standard three-tier architecture and adopts the designing idea of MVC pattern with some advanced technologies, such as Saliency detection, Meanshift clusting, Possion blending etc.When using the system, first, the user can choose an existing hand-drawn sketch or create a new sketch, and provides the keywords of the foreground and background according to the sketch. Then, the user can choose system usage pattern in online or offline mode. Next, it only needs to click the button segmentation, filtering, and synthesis consecutive, the system will complete the processing. The processing is hidden to the user. Finally, the system shows some synthetic images according to user’s requirements, which mean the processing complete. The advantages of this system is that users don’t need to provide original synthetic images material and can create more composite images in fewer interactions, and it provides two kinds of operating mode, online mode and offline mode.