SPP-Net算法的相关论文
SPP-Net的第一作者也是何凯明，原论文《Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition》。用于分类和检测任务，在ImageNet数据集ILSVRC2014竞赛上，检测任务获得第二名、分类任务第三名。
Abstract
Existing deep convolutional neural networks (CNNs) require a fixed-size (e.g., 224×224) input image. This requirement is “artificial” and may reduce the recognition accuracy for the images or sub-images of an arbitrary size/scale. In this work, we equip the networks with another pooling strategy, “spatial pyramid pooling”, to eliminate the above requirement. The new network structure, called SPP-net, can generate a fixed-length representation regardless of image size/scale. Pyramid pooling is also robust to object deformations. With these advantages, SPP-net should in general improve all CNN-based image classification methods. On the ImageNet 2012 dataset, we demonstrate that SPP-net boosts the accuracy of a variety of CNN architectures despite their different designs. On the Pascal VOC 2007 and Caltech101 datasets, SPP-net achieves state-of-theart classification results using a single full-image representation and no fine-tuning.
现有的深度卷积神经网络(CNNs)需要一个固定大小的输入图像(如224×224)。这一要求是“人为的”，可能会降低对任意大小/尺度的图像或子图像的识别精度。在这项工作中，我们为网络配备了另一种pooling 策略，“空间金字塔池”，以消除上述的要求。这种新的网络结构称为SPP-net，可以生成固定长度的表示，而不受图像大小/比例的影响。金字塔池对物体变形也有很强的鲁棒性。基于这些优点，SPP-net一般应改进所有基于CNN的图像分类方法。在ImageNet 2012数据集中，尽管它们的设计不同，我们证明了SPP-net提高了各种CNN架构的准确性。在Pascal VOC 2007和Caltech101数据集上，SPP-net使用单一的全图像表示，无需微调，就可以实现最先进的分类结果。
The power of SPP-net is also significant in object detection. Using SPP-net, we compute the feature maps from the entire image only once, and then pool features in arbitrary regions (sub-images) to generate fixed-length representations for training the detectors. This method avoids repeatedly computing the convolutional features. In processing test images, our method is 24-102× faster than the R-CNN method, while achieving better or comparable accuracy on Pascal VOC 2007.
在目标检测中，SPP-net的能力也很重要。利用SPP-net算法，只对整个图像进行一次特征映射计算，然后将特征集合到任意区域(子图像)，生成固定长度的表示形式，用于训练检测器。该方法避免了卷积特征的重复计算。在处理测试图像时，我们的方法比R-CNN方法快24-102倍，而在Pascal VOC 2007上达到了更好或相近的精度。
In ImageNet Large Scale Visual Recognition Challenge (ILSVRC) 2014, our methods rank #2 in object detection and #3 in image classification among all 38 teams. This manuscript also introduces the improvement made for this competition.
在2014年的ImageNet Large Scale Visual Recognition Challenge (ILSVRC)中，我们的方法在所有38个团队中对象检测排名第二，图像分类排名第三。本文还介绍了本次比赛的改进情况。
CONCLUSION
SPP is a flexible solution for handling different scales, sizes, and aspect ratios. These issues are important in visual recognition, but received little consideration in the context of deep networks. We have suggested a solution to train a deep network with a spatial pyramid pooling layer. The resulting SPP-net shows outstanding accuracy in classification/detection tasks and greatly accelerates DNN-based detection. Our studies also show that many time-proven techniques/insights in computer vision can still play important roles in deep-networks-based recognition.
结论
SPP是一个灵活的解决方案，可以处理不同的规模、大小和纵横比。这些问题在视觉识别中很重要，但在深度网络环境中却很少被考虑。论文提出了一种利用空间金字塔池层，训练深度网络的方法。由此产生的SPP-net在分类/检测任务中显示出优异的精度，大大加快了基于DNN的检测速度。我们的研究还表明，在基于深度网络的识别中，许多经过时间检验的计算机视觉技术/见解仍然可以发挥重要作用。
相关论文
Kaiming He, XiangyuZhang, ShaoqingRen, and Jian Sun.
Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition . ECCV 2014
https://arxiv.org/abs/1406.4729
0、实验结果
1、VOC2007
2、ILSVRC 2014 Classification
3、ILSVRC 2014 Detection
1、SPP-Net中的亮点
在此之前，所有的神经网络都是需要输入固定尺寸的图片，比如224*224（ImageNet）、32*32(LenNet)、96*96等。这样对于我们希望检测各种大小的图片的时候，需要经过crop，或者warp等一系列操作，这都在一定程度上导致图片信息的丢失和变形，限制了识别精确度。而且，从生理学角度出发，人眼看到一个图片时，大脑会首先认为这是一个整体，而不会进行crop和warp，所以更有可能的是，我们的大脑通过搜集一些浅层的信息，在更深层才识别出这些任意形状的目标。
分类: improves all CNN architectures
检测: 24~64x faster than R-CNN
ILSVRC 2014: #2 in detection, #3 in classification.
