Video Content Expression Oriented Comics Generation System

Comics,as a popular art form,is popular among people all over the world.However,the threshold of cartoon creation is very high,which requires the creators to have high painting skills,composition ability and artistic accomplishment.At the same time,comic creation is a time-consuming task.A common page of a comic book often requires the creator to spend hours or even days to create.Although there are already a lot of computer software tools to assist creators in creating,comic creation is still a time-consuming,labor-intensive and high-threshold job.Therefore,in order to help ordinary people to create comics and improve the efficiency of professional creators,we propose a comics generation system that uses ordinary video as the creative material for video content expression.The system can converts the input video material into a comic book without human intervention.Meanwhile,in order to generate better results,we allow manual intervention and fine-tuning the results in each subprocess.The system mainly contains the following four modules: key frames selection and extraction,image stylization,automatic layout and automatic balloon generation and placement.Firstly,we need to select and extract the key frames from the input video as he basic material of the comic generation.The selected key frames are the most critical,representative and most attractive frames of the input video.Secondly,we perform comic stylization on the extracted key frames,aiming to convert ordinary images to comic-like images.Thirdly,we need to layout the stylized images automatically.Automatic layout contains two main tasks: assigning comic images to different pages and cropping,morphing,and placing images on the same page.Finally,we use the dialog content in the video to automatically generate balloon boxes of different shapes and place the balloon box in the right place.Meanwhile,in order to help users to use our system,we have developed a software interface,through which users can not only automatically generate comics from the video,but also can manually adjust it to get the results they want.