TSingNet: Scale-aware and context-rich feature learning for traffic sign detection and recognition in the wild
TSingNet: Scale-aware and context-rich feature learning for traffic sign detection and recognition in the wild
复制标题
TSingNet：用于野外交通标志检测和识别的尺度感知和上下文丰富的特征学习
DOI：
10.1016/j.neucom.2021.03.049
复制
发表时间：
2021-04-10
期刊：
影响因子：
6
通讯作者：
Fu, Zhang-Hua
中科院分区：
文献类型：
--
作者：
Liu, Yuanyuan;Peng, Jiyao;Fu, Zhang-Hua
野外交通标志检测与识别是一项具有挑战性的任务。由于尺度变化和上下文丢失，现有技术往往无法检测到小型或被遮挡的交通标志，这导致了多尺度之间的语义差距。我们提出了一种新的交通标志检测网络（TSingNet），它学习具有尺度感知和丰富上下文的特征，以有效地检测和识别野外的小型和被遮挡的交通标志。具体而言，TSingNet首先构建了一个注意力驱动的双边特征金字塔网络，该网络借鉴了自下而上和自上而下的子网，在尺度自注意力学习中双重循环低、中、高层次的前景语义。这是为了学习具有尺度感知的前景特征，从而缩小多尺度之间的语义差距。然后引入了一个具有可变膨胀率的自适应感受野融合块，以利用丰富的上下文表示并抑制每个尺度上遮挡的影响。TSingNet通过联合最小化尺度感知损失和多分支融合损失进行端到端的训练，这增加了少量参数，但显著提高了检测性能。在对三个具有挑战性的交通标志数据集（TT100K、STSD和DFG）进行的大量实验中，TSingNet在野外交通标志检测和识别方面优于最先进的方法。（c）2021年由爱思唯尔出版公司出版。
Traffic sign detection and recognition in the wild is a challenging task. Existing techniques are often incapable of detecting small or occluded traffic signs because of the scale variation and context loss, which causes semantic gaps between multiple scales. We propose a new traffic sign detection network (TSingNet), which learns scale-aware and context-rich features to effectively detect and recognize small and occluded traffic signs in the wild. Specifically, TSingNet first constructs an attention-driven bilateral feature pyramid network, which draws on both bottom-up and top-down subnets to dually circulate low-, mid-, and high-level foreground semantics in scale self-attention learning. This is to learn scale aware foreground features and thus narrow down the semantic gaps between multiple scales. An adaptive receptive field fusion block with variable dilation rates is then introduced to exploit context-rich representation and suppress the influence of occlusion at each scale. TSingNet is end-to-end trainable by joint minimization of the scale-aware loss and multi-branch fusion losses, this adds a few parameters but significantly improves the detection performance. In extensive experiments with three challenging traffic sign datasets (TT100K, STSD and DFG), TSingNet outperformed state-of-the-art methods for traffic sign detection and recognition in the wild.& nbsp; (c) 2021 Published by Elsevier B.V.