[论文解读] Object Detection with Deep Learning: A Review

Zhong‐Qiu Zhao, Peng Zheng+2 more|arXiv (Cornell University)|Jul 15, 2018

Visual Attention and Saliency Detection参考文献 173被引用 82

一句话总结

基于深度学习的对象检测的综合综述，追踪从 R-CNN 到现代区域基于和回归/分类框架，并在通用和特定任务上进行分析。

ABSTRACT

Due to object detection's close relationship with video analysis and image understanding, it has attracted much research attention in recent years. Traditional object detection methods are built on handcrafted features and shallow trainable architectures. Their performance easily stagnates by constructing complex ensembles which combine multiple low-level image features with high-level context from object detectors and scene classifiers. With the rapid development in deep learning, more powerful tools, which are able to learn semantic, high-level, deeper features, are introduced to address the problems existing in traditional architectures. These models behave differently in network architecture, training strategy and optimization function, etc. In this paper, we provide a review on deep learning based object detection frameworks. Our review begins with a brief introduction on the history of deep learning and its representative tool, namely Convolutional Neural Network (CNN). Then we focus on typical generic object detection architectures along with some modifications and useful tricks to improve detection performance further. As distinct specific detection tasks exhibit different characteristics, we also briefly survey several specific tasks, including salient object detection, face detection and pedestrian detection. Experimental analyses are also provided to compare various methods and draw some meaningful conclusions. Finally, several promising directions and tasks are provided to serve as guidelines for future work in both object detection and relevant neural network based learning systems.

研究动机与目标

• 综述与对象检测相关的深度学习与 CNN 的历史与基础。
• 总结通用对象检测架构及其从基于区域提案到回归/分类框架的演变。
• 评审显著性对象检测、人脸检测和行人检测等应用并比较性能见解。
• 突出早期方法的局限性并展示随后的改进与实际考虑。

提出的方法

• 回顾并将对象检测框架分为基于区域提案的和基于回归/分类的管线。
• 讨论关键架构（R-CNN、SPP-net、Fast R-CNN、Faster R-CNN、R-FCN、FPN、Mask R-CNN）及其训练策略。
• 解释 Region Proposal Network (RPN) 及其在端到端训练和速度提升中的作用。
• 描述提高效率和精度的改进，包括 RoI pooling/align、多任务损失和特征金字塔（FPN）。
• 概述在显著性、人脸和行人检测方面的特定任务适应。

实验结果

研究问题

• RQ1 基于深度学习的对象检测的主要架构范式有哪些，它们如何随时间演变？
• RQ2 基于区域提案的方法在准确性和效率方面与回归/分类方法相比如何？
• RQ3 有哪些关键技术将实时端到端对象检测拉近？
• RQ4 专门任务（显著对象检测、人脸、行人）如何影响检测器设计与性能？

主要发现

• R-CNN 在 PASCAL VOC 2012 上达到 mean average precision (mAP) 53.3%，显著优于早期方法。
• SPP-net 和 Fast R-CNN 通过重复使用卷积特征图并引入带多任务损失的 RoI 池化，提高了效率和精度。
• Faster R-CNN 引入 Region Proposal Networks (RPN) 以端到端生成提案，使训练更快、集成度更高，并实现接近实时的性能（例如在 VOC07/12 上的 GPU 5 FPS）。
• R-FCN 提出了一种全卷积架构，带位置敏感 RoI 池化，以平衡平移变异性和效率，在 PASCAL VOC 和 COCO 上取得了不错的结果。
• FPN 引入了自顶向下的横向连接特征金字塔，以统一多尺度特征表示，提升跨尺度的检测。
• Mask R-CNN 将 Faster R-CNN 扩展为独立的分割分支和 RoIAlign，以在不牺牲检测性能的情况下提高掩膜精度。