Jiawei Wang
ABSTRACT
The complex background, uneven illumination and object occlusion have increased the difficulty of scene texts detection. In this paper, we improved the existing object detection algorithm SSD, and made it possible to detect text objects in traffic guidance sign. We used a deep neural network CRNN to identify the text. This network is a combination of Convolution Neural Network and Recurrent Neural Network. At the same time, we proposed a new idea to optimize the detection algorithm through the text recognition result, so that the whole network can be trained end-to-end. According to the experimental results, the detection network achieves 88% mAP on our dataset at 11.6FPS, which has a good recognition effect.
- HUANG H, HOU L Y. 2017. Speed limit sign detection based on Gaussian color model and template matching. C. IEEE International Conference on Vision, Image and Signal Processing, 118--122.Google Scholar
- XU Z, REN J Y, BAO C Q, 2016. A Traffic signs' detection method of contour approximation based on concave removal. C. 2016 Chinese Control and Decision Conference, 5199--5204.Google Scholar
- WANG C Y. 2018. Research and application of traffic sign detection and recognition based on deep learning. C. International Conference on Robots & Intelligent System, 150--152.Google Scholar
- LEE H S, KIM K. 2018. Simultaneous traffic sign detection and boundary estimation using convolutional neural net- work. J. IEEE Transactions on Intelligent Transportation Systems, 19( 5): 1652--1663.Google ScholarDigital Library
- NATARAJAN S, ANNAMRAJU A, BARADKAR C S. 2018. Traffic sign recognition using weighted multi-convolutional neural network. J. IET Intelligent Transport Systems, 12(10): 1396--1405.Google ScholarCross Ref
- Shi B, Bai X, Yao C. 2017. An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition. J. IEEE Transactions on Pattern Analysis & Machine Intelligence, PP(99): 1--1.Google ScholarDigital Library
- Tian S, Yin X C, Su Y, et al. 2017. A Unified Framework for Tracking Based Text Detection and Recognition from Web Videos. J. IEEE Transactions on Pattern Analysis & Machine Intelligence, PP(99): 542--554.Google Scholar
- Simonyan K, Zisserman A. 2014. Very Deep Convolutional Networks for Large-Scale Image Recognition. J. Computer Science.Google Scholar
- Gupta A, Vedaldi A, Zisserman A. 2016. Synthetic Data for Text Localisation in Natural Images. J. 2315--2324.Google Scholar
- Ren, SQ, He, KM, Girshick, R, Sun, J. 2017. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. J. IEEE transactions on pattern analysis and machine intelligence, 39(6):1137--1149.Google Scholar
- Diederik P. Kingma, Jimmy Ba. 2015. Adam: A Method for Stochastic Optimization. J. The 3rd International Conference for Learning Representations.Google Scholar
- Zhou X, Yao C, Wen H, et al. 2017. EAST: An Efficient and Accurate Scene Text Detector. J. IEEE Conference on Computer Vision and Pattern Recognition, 2642--2651.Google Scholar
Index Terms
- Street Sign Recognition Algorithm Based on Deep Learning
Recommendations
Traffic Sign Recognition Based on Deep Learning Technique
ICISDM '22: Proceedings of the 6th International Conference on Information System and Data MiningTraffic sign recognition plays a significant role in intelligent transportation system. Therefore, in this paper, I propose a traffic sign recognition algorithm based on Convolutional Neural Network (CNN). The dataset collected to train and test in ...
Deep neural network for traffic sign recognition systems: An analysis of spatial transformers and stochastic optimisation methods
AbstractThis paper presents a Deep Learning approach for traffic sign recognition systems. Several classification experiments are conducted over publicly available traffic sign datasets from Germany and Belgium using a Deep Neural Network ...
Highlights- A Deep Neural Network that is top-1 ranked in the German traffic sign benchmark.
Exploiting synergies of mobile mapping sensors and deep learning for traffic sign recognition systems
A traffic sign recognition system based on 3D point clouds and deep learning.A 3D point cloud processing workflow aiming to detect reflective traffic sign panels.Proposed deep neural network gets top-1 rank on German traffic sign benchmark.A Spanish ...
Comments