选择对象检测方法的标准
Criteria to choose a object detection Method
我正处于项目的研究阶段，我正在尝试使用 CNN 制作对象检测器。我知道通常有 2 种“类型”的 CNN 对象检测器，基于区域建议（即 R-CNN 和 R-FCN ）和基于 Regression/Classification 的方法（即 YOLO 和 SSD）。问题是我不太确定应该使用哪种方法。我想知道选择一种方法而不是另一种方法的通常理由是什么。有一些一般标准，例如速度与准确性。但是还有其他常用的推理吗？
探测器有两类，一级和二级。 Yolo、SSD、RetinaNet、CenterNet等属于一个阶段，而R-FCN、R-CNN、Faster R-CNN等属于两个阶段。
直接引用 [1] 关于两级检测器与一级相比的优势，
Compared to one-stage detectors, the two-stage ones have the following
advantages: 1) By sampling a sparse set of region proposals, two-stage
detectors filter out most of the negative proposals; while one-stage
detectors directly face all the regions on the image and have a
problem of class imbalance if no specialized design is introduced. 2)
Since two-stage detectors only process a small number of proposals,
the head of the network (for proposal classification and regression)
can be larger than one-stage detectors, so that richer features will
be extracted. 3) Two-stage detectors have high-quality features of
sampled proposals by use of the RoIAlign [10] operation that extracts
the location consistent feature of each proposal; but different region
proposals can share the same feature in one-stage detectors and the
coarse and spatially implicit representation of proposals may cause
severe feature misalignment. 4) Two-stage detectors regress the object
location twice (once on each stage) and the bounding boxes are better
refined than one-stage methods.
报价准确性与效率，
One-stage detectors are more efficient and elegant in design, but
currently the two-stage detectors have domination in accuracy.
可以将一级检测器部署在手机等边缘设备上，以进行快速实时检测。与更多计算密集型检测器相比，这可以节省更多能量。
总而言之，如果精度更重要，则选择两级检测器，否则选择一级检测器以进行更快的检测，同时保持足够好的精度。
[1] 的相关作品部分包含易于阅读的详细信息，每篇参考论文都有两个阶段和一个阶段的评论。
对象检测基准
https://paperswithcode.com/task/object-detection
参考资料
[1] 模拟数据，https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123590528.pdf
[2] Speed/accuracy 现代卷积对象检测器的权衡，https://arxiv.org/pdf/1611.10012.pdf
[3] 视网膜网络，https://arxiv.org/pdf/1708.02002.pdf
[4] 对象检测审查，https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9186021
[5] CSPNET，https://arxiv.org/pdf/1911.11929v1.pdf
[6] 中心网，https://arxiv.org/pdf/1904.08189v3.pdf
[7] EfficientDet，https://arxiv.org/pdf/1911.09070.pdf
[8] 脊柱网络，https://arxiv.org/pdf/1912.05027.pdf
相关文章
我正处于项目的研究阶段，我正在尝试使用 CNN 制作对象检测器。我知道通常有 2 种“类型”的 CNN 对象检测器，基于区域建议（即 R-CNN 和 R-FCN ）和基于 Regression/Classification 的方法（即 YOLO 和 SSD）。问题是我不太确定应该使用哪种方法。我想知道选择一种方法而不是另一种方法的通常理由是什么。有一些一般标准，例如速度与准确性。但是还有其他常用的推理吗？
探测器有两类，一级和二级。 Yolo、SSD、RetinaNet、CenterNet等属于一个阶段，而R-FCN、R-CNN、Faster R-CNN等属于两个阶段。
直接引用 [1] 关于两级检测器与一级相比的优势，
Compared to one-stage detectors, the two-stage ones have the following advantages: 1) By sampling a sparse set of region proposals, two-stage detectors filter out most of the negative proposals; while one-stage detectors directly face all the regions on the image and have a problem of class imbalance if no specialized design is introduced. 2) Since two-stage detectors only process a small number of proposals, the head of the network (for proposal classification and regression) can be larger than one-stage detectors, so that richer features will be extracted. 3) Two-stage detectors have high-quality features of sampled proposals by use of the RoIAlign [10] operation that extracts the location consistent feature of each proposal; but different region proposals can share the same feature in one-stage detectors and the coarse and spatially implicit representation of proposals may cause severe feature misalignment. 4) Two-stage detectors regress the object location twice (once on each stage) and the bounding boxes are better refined than one-stage methods.
报价准确性与效率，
One-stage detectors are more efficient and elegant in design, but currently the two-stage detectors have domination in accuracy.
可以将一级检测器部署在手机等边缘设备上，以进行快速实时检测。与更多计算密集型检测器相比，这可以节省更多能量。
总而言之，如果精度更重要，则选择两级检测器，否则选择一级检测器以进行更快的检测，同时保持足够好的精度。
[1] 的相关作品部分包含易于阅读的详细信息，每篇参考论文都有两个阶段和一个阶段的评论。
对象检测基准
https://paperswithcode.com/task/object-detection
参考资料
[1] 模拟数据，https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123590528.pdf
[2] Speed/accuracy 现代卷积对象检测器的权衡，https://arxiv.org/pdf/1611.10012.pdf
[3] 视网膜网络，https://arxiv.org/pdf/1708.02002.pdf
[4] 对象检测审查，https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9186021
[5] CSPNET，https://arxiv.org/pdf/1911.11929v1.pdf
[6] 中心网，https://arxiv.org/pdf/1904.08189v3.pdf
[7] EfficientDet，https://arxiv.org/pdf/1911.09070.pdf
[8] 脊柱网络，https://arxiv.org/pdf/1912.05027.pdf