[论文解读] Improved detection of small objects in road network sequences

Iván García, Rafael Marcos Luque‐Baena+1 more|arXiv (Cornell University)|May 18, 2021

Advanced Image Processing Techniques被引用 2

一句话总结

本文提出一种两阶段目标检测框架，通过在预训练卷积神经网络识别出的感兴趣区域上应用超分辨率技术，显著提升道路网络监控视频中小目标的检测性能。通过放大检测到的小目标并重新处理，该方法在不微调基础模型的前提下，显著提高了平均精度（mAP），尤其在小尺寸车辆检测方面，某些情况下检测率最高提升2.3倍。

ABSTRACT

The vast number of existing IP cameras in current road networks is an opportunity to take advantage of the captured data and analyze the video and detect any significant events. For this purpose, it is necessary to detect moving vehicles, a task that was carried out using classical artificial vision techniques until a few years ago. Nowadays, significant improvements have been obtained by deep learning networks. Still, object detection is considered one of the leading open issues within computer vision. The current scenario is constantly evolving, and new models and techniques are appearing trying to improve this field. In particular, new problems and drawbacks appear regarding detecting small objects, which correspond mainly to the vehicles that appear in the road scenes. All this means that new solutions that try to improve the low detection rate of small elements are essential. Among the different emerging research lines, this work focuses on the detection of small objects. In particular, our proposal aims to vehicle detection from images captured by video surveillance cameras. In this work, we propose a new procedure for detecting small-scale objects by applying super-resolution processes based on detections performed by convolutional neural networks \emph{(CNN)}. The neural network is integrated with processes that are in charge of increasing the resolution of the images to improve the object detection performance. This solution has been tested for a set of traffic images containing elements of different scales to test the efficiency according to the detections obtained by the model, thus demonstrating that our proposal achieves good results in a wide range of situations.

研究动机与目标

• 解决道路网络视频序列中目标检测对小目标精度低的长期挑战。
• 在不微调预训练模型的前提下，提升低分辨率、高杂波交通场景中小车辆的目标检测性能。
• 通过增强候选区域的空间分辨率，缓解深层网络因下采样导致的特征损失。
• 通过迭代式超分辨率与重检测，提升小目标检测的可靠性与召回率。
• 在计算开销极小的前提下，实现对现有交通监控系统的实用化部署。

提出的方法

• 使用预训练的目标检测模型（如EfficientDet、CenterNet）对输入视频帧中的所有目标进行检测。
• 从检测到的目标中心裁剪出区域，尤其针对小目标进行分辨率增强。
• 应用深度超分辨率网络（如EDSR或类似模型）将每个检测区域放大至更高分辨率。
• 使用相同的预训练检测器在超分辨率后的区域中重新检测目标，以优化预测结果。
• 聚合原始检测与增强后检测的结果，以提升整体检测置信度与召回率。
• 采用COCO评估指标（IoU阈值下的mAP）对增强前后性能进行定量比较。

实验结果

研究问题

• RQ1 对检测到的目标区域进行超分辨率处理，能否提升交通监控视频中小车辆的检测精度？
• RQ2 两阶段检测方法——先检测后超分辨率再重检测——是否相比单阶段推理获得更高的mAP？
• RQ3 该方法在多大程度上恢复了因分辨率损失而被基础检测器遗漏的小目标？
• RQ4 在真实世界交通序列中，性能提升在不同目标尺度（小、中、大）上的表现如何？
• RQ5 该方法是否可直接应用于现有预训练模型而无需微调，从而实现在实际交通监控系统中的部署？

主要发现

• 与基础模型相比，该方法在IoU=0.50:0.95下，小目标的mAP最高提升2.3倍，其中Video 1的提升最为显著（从0.168提升至0.298）。
• 在Video 2中，小目标的mAP由EfficientDet D5的0.158提升至0.207，相对提升31%。
• 在Video 3中，小目标的mAP由EfficientDet D5的0.056提升至0.082，相对提升46%，表明在多样化场景中均保持稳定增益。
• 视觉对比与检测数量图（图13–15）确认，该方法检测到了首次检测中未识别出的额外小车辆。
• 该方法在所有IoU和面积阈值下均实现了更高的mAP，尤其在小目标检测中增益最为显著，验证了其在低分辨率交通监控中的有效性。
• 该方法无需微调基础检测器即可保持高性能，可高效集成至现有交通监控流水线。