Visual Recognition of Traffic Signs in Natural Scenes Based on Improved RetinaNet

Aiming at recognizing small proportion, blurred and complex traffic sign in natural scenes, a traffic sign detection method based on RetinaNet-NeXt is proposed. First, to ensure the quality of dataset, the data were cleaned and enhanced to denoise. Secondly, a novel backbone network ResNeXt was employed to improve the detection accuracy and effection of RetinaNet. Finally, transfer learning and group normalization were adopted to accelerate our network training. Experimental results show that the precision, recall and mAP of our method, compared with the original RetinaNet, are improved by 9.08%, 9.09% and 7.32%, respectively. Our method can be effectively applied to traffic sign detection.

Improved YOLO v5 with balanced feature pyramid and attention module for traffic sign detection

With the development of automatic driving technology, traffic sign detection has become a very important task. However, it is a challenging task because of the complex traffic sign scene and the small size of the target. In recent years, a number of convolutional neural network (CNN) based object detection methods have brought great progress to traffic sign detection. Considering the still high false detection rate, as well as the high time overhead and computational overhead, the effect is not satisfactory. Therefore, we employ lightweight network model YOLO v5 (You Only Look Once) as our work foundation. In this paper, we propose an improved YOLO v5 method by using balances feature pyramid structure and global context block to enhance the ability of feature fusion and feature extraction. To verify our proposed method, we have conducted a lot of comparative experiments on the challenging dataset Tsinghua-Tencent-100K (TT100K). The experimental results demonstrate that the [email protected] and [email protected]:0.95 are improved by 1.9% and 2.1%, respectively.

Traffic Sign Detection and Recognition using Image Processing

Now a days, automobiles became most convenient mode of transportation for everyone. As we know one of the most important functions, TSDR has become a popular research . It primarily involves the use of vehicle cameras to collect real- time road pictures and then recognize and identify traffic signs seen on the road, therefore delivering correct data to the driving system. With the advancement of science and technology, an increasing number of scholars are turning to deep learning technology to save time in traditional processes. From the training samples, this model can learn the deep features inside the autonomously. The accuracy and great efficiency of detection and identification are the subject of this essay. A deep convolution neural network algorithm is proposed to train traffic sign training sets using Caffe[3], an open-source framework, in order to obtain a model that can classify traffic signs and learn and identify the most critical of these traffic sign features, in order to achieve the goal of identifying traffic signs in the real world. Keywords: Traffic sign, Segmentation, Gabor filter, Traffic Sign Detection and Recognition (TSDR)

Traffic Sign Detection Using Region And Corner Feature Extraction Method

Traffic signs are an important feature in providing safety information for drivers about road conditions. Recognition of traffic signs can reduce the burden on drivers remembering signs and improve safety. One solution that can reduce these violations is by building a system that can recognize traffic signs as reminders to motorists. The process applied to traffic sign detection is image processing. Image processing is an image processing and analysis process that involves a lot of visual perception. Traffic signs can be detected and recognized visually by using a camera as a medium for retrieving information from a traffic sign. The layout of different traffic signs can affect the identification process. Several studies related to the detection and recognition of traffic signs have been carried out before, one of the problems that arises is the difficulty in knowing the kinds of traffic signs. This study proposes a combination of region and corner point feature extraction methods. Based on the test results obtained an accuracy value of 76.2%, a precision of 67.3 and a recall value of 78.6.

REAL OBJECT DETECTION SYSTEM FOR IRAQ TRAFFIC SIGNS BASED ON MASK R-CNN

Traffic signs object detection has gained great interest in recent years, as one of the most important object detector applications. Traffic signs detection is based on deep learning, which gives it the benefit of high detection precision and timely response to condition changes of the traffic. Therefore, this paper shows an efficient method for detecting traffic signs in real-time. Hence, it implements a new Iraqi Traffic Sign Detection Benchmark (IQTSDB) dataset based on Mask Region-based Convolutional Neural Network (Mask R-CNN). The results show that the implementation of IQTSDB dataset with Mask R-CNN has a great efficiency in different conditions such as sunny, cloudy, weak light, and rainy conditions. In addition, the real video captured for traffic signs in Baghdad has been taken and compared to the German Traffic Signs Detection Benchmark (GTSDB) dataset. The IQTSDB dataset has a better performance than GTSDB dataset based on the performance parameters training loss and mean Average Precision (mAP).

Detection of Small Size Traffic Signs Using Regressive Anchor Box Selection and DBL Layer Tweaking in YOLOv3

Traffic sign recognition is a key module of autonomous cars and driver assistance systems. Traffic sign detection accuracy and inference time are the two most important parameters. Current methods for traffic sign recognition are very accurate; however, they do not meet the requirement for real-time detection. While some are fast enough for real-time traffic sign detection, they fall short in accuracy. This paper proposes an accuracy improvement in the YOLOv3 network, which is a very fast detection framework. The proposed method contributes to the accurate detection of a small-sized traffic sign in terms of image size and helps to reduce false positives and miss rates. In addition, we propose an anchor frame selection algorithm that helps in achieving the optimal size and scale of the anchor frame. Therefore, the proposed method supports the detection of a small traffic sign with real-time detection. This ultimately helps to achieve an optimal balance between accuracy and inference time. The proposed network is evaluated on two publicly available datasets, namely the German Traffic Sign Detection Benchmark (GTSDB) and the Swedish Traffic Sign dataset (STS), and its performance showed that the proposed approach achieves a decent balance between mAP and inference time.