SReN: Shape regression network for comic storyboard extraction

Abstract

The goal of storyboard extraction is to decompose the comic image into storyboards, which is the fundamental step of comic image understanding and producing digital comic documents suitable for mobile reading. Most of existing approaches are based on hand crafted low-level visual patters like edge segments and line segments, which do not capture high-level vision information. To overcome this drawback of the existing approaches, we propose a novel architecture based on deep convolutional neural network, named Shape Regression Network (SReN), to detect storyboards within comic images. Firstly, we use Fast R-CNN to generate rectangle bounding boxes as storyboard proposals. Then we train a deep neural network to predict quadrangles for these proposals. Unlike existing object detection methods which only output rectangle bounding boxes, SReN can produce more precise quadrangle bounding boxes. Experimental results on 7382 comic pages, demonstrate that SReN outperforms the state-of-the-art methods by more than 10% in terms of F1-score and page correction rate.