A Multiscale Recognition Method for the Optimization of Traffic Signs Using GMM and Category Quality Focal Loss

Abstract

Effective traffic sign recognition algorithms can assist drivers or automatic driving systems in detecting and recognizing traffic signs in real-time. This paper proposes a multiscale recognition method for traffic signs based on the Gaussian Mixture Model (GMM) and Category Quality Focal Loss (CQFL) to enhance recognition speed and recognition accuracy. Specifically, GMM is utilized to cluster the prior anchors, which are in favor of reducing the clustering error. Meanwhile, considering the most common issue in supervised learning (i.e., the imbalance of data set categories), the category proportion factor is introduced into Quality Focal Loss, which is referred to as CQFL. Furthermore, a five-scale recognition network with a prior anchor allocation strategy is designed for small target objects i.e., traffic sign recognition. Combining five existing tricks, the best speed and accuracy tradeoff on our data set (40.1% mAP and 15 FPS on a single 1080Ti GPU), can be achieved. The experimental results demonstrate that the proposed method is superior to the existing mainstream algorithms, in terms of recognition accuracy and recognition speed.

A Novel Neural Network Model for Traffic Sign Detection and Recognition under Extreme Conditions

“…Traditional Traffic Sign Recognition Technology. The traditional recognition algorithms are aimed at locating the region of interest and identifying the classification [1]. In the study by Zhou and Deng [24], color and spaces were combined so that the traffic sign colors can be treated as one class.…”

“…Under severe weather conditions such as fog and snowy, traffic accidents occur frequently due to distracted driving, inattentiveness, or poor visibility [1]. To decrease the risk of accidents and improve the driving experience of drivers, traffic sign recognition systems (TSRS) have been developed and played an important role in autonomous driving and road network maintenance [2].…”

“…(i) Under different weather conditions such as foggy, snowy, and rainy days, the captured traffic sign images may contain a lot of noises and distortions [6] (ii) Under different viewing angles and viewing distance, the traffic sign images may appear distortions of shape and color [6] (iii) The complicated road environment can lead to complex background of the traffic signs [1] (iv) Under different light conditions, color distortion or saturation may impact the quality of captured images [1] (v) Traffic signs may be partially hidden by trees, snow in some conditions…”

“…The most representative models are YOLO [13][14][15][16][17][18], SSD [19], and RetinaNet [20]. The one-stage traffic sign recognition model has high efficiency but insufficient accuracy [1].…”

Multiscale Traffic Sign Detection Method in Complex Environment Based on YOLOv4

“…These algorithms can automatically locate and identify traffic signs, significantly improving the recognition speed [ 6 ]. However, traffic sign detection and recognition still face the following challenges: In rainy, snowy, foggy, and other complicated weather, the photos captured by the camera contain a significant amount of noise [ 7 ] Under different lighting conditions, the color and saturation of traffic signs will change [ 8 ] Some part of the traffic sign may be blocked by railings, trees, snow, etc. Under different shooting angles, the shape of traffic signs may be distorted Some traffic signs are too small to be recognized …”