Scene Text Detection and Segmentation based on Cascaded Convolution Neural Networks

Abstract

Scene text detection and segmentation are two important and challenging research problems in the field of computer vision. This paper proposes a novel method for scene text detection and segmentation based on cascaded convolution neural networks (CNNs). In this method, a CNN based text-aware candidate text region (CTR) extraction model (named detection network, DNet) is designed and trained using both the edges and the whole regions of text, with which coarse CTRs are detected. A CNN based CTR refinement model (named segmentation network, SNet) is then constructed to precisely segment the coarse CTRs into text to get the refined CTRs. With DNet and SNet, much fewer CTRs are extracted than with traditional approaches while more true text regions are kept. The refined CTRs are finally classified using a CNN based CTR classification model (named classification network, CNet) to get the final text regions.

Existing System

Sliding window based methods and stroke width transform (SWT), maximally stable extremal region (MSER)

Proposed System

This paper proposes a novel method for scene text detection and segmentation based on cascaded convolution neural networks. In this method, a text-aware CTR extraction model and a CTR refinement model are devised to extract CTRs and obtain precise text segmentation results, which can overcome the above problems. The text-aware CTR extraction model detects regions of text (or the coarse CTRs) in the scene images and the CTR refinement model precisely segments the detected regions (or the coarse CTRs) into text in order to get the refined CTRs. Finally, the refined CTRs are fed into a CTR classification model to filter out non-text regions and obtain the final text regions. All of these models are based on a powerful CNN model.

Architecture

BLOCK DIAGRAM