CASCADED SEGMENTATION-DETECTION NETWORKS FOR TEXT-BASED TRAFFIC SIGN DETECTION

Abstract

In this paper, we propose a novel text-based traffic sign detection framework with two deep learning components. More precisely, we apply a fully convolutional network to segment candidate traffic sign areas providing candidate regions of interest (RoI), followed by a fast neural network to detect texts on the extracted RoI. The proposed method makes full use of the characteristics of traffic signs to improve the efficiency and accuracy of text detection. On one hand, the proposed two-stage detection method reduces the search area of text detection and removes texts outside traffic signs. On the other hand, it solves the problem of multi-scales for the text detection part to a large extent. Extensive experimental results show that the proposed method achieves the state-of-the-art results on the publicly available traffic sign data set: Traffic Guide Panel data set. In addition, we collect a data set of text-based traffic signs including Chinese and English traffic signs. Our method also performs well on this data set, which demonstrates that the proposed method is general in detecting traffic signs of different languages.

Existing System

A fast text detector with a single deep neural network

Proposed System

The main contributions of this paper are summarized as below: 1) A new cascaded segmentation-detection framework which can be trained for text-based traffic sign detection is proposed; 2) The proposed method makes full use of the information of traffic signs to help text detection process, which improves the detection speed, removes the background texts, and solves the problem of multi-scale for text detection; 3) The proposed method is simple and effective, which outperforms the state-of-the-art methods in terms of speed and accuracy. On the dataset used in, the proposed method achieves an F-measure (88%), which is 20% higher than the previous method proposed in, and meets the real-time requirements; 4) A new dataset of Text-based traffic signs in English and Chinese is collected. The experiment results prove that the proposed method can be easily extended to text-based traffic sign detection in other languages.

Architecture

BLOCK DIAGRAM

ARCHITECTURE