在数学(特别是功能分析)中,卷积是对两个函数(f和g)的数学运算,产生三个函数,表示第一个函数的形状如何被另一个函数修改。 卷积一词既指结果函数,又指计算结果的过程。 它定义为两个函数的乘积在一个函数反转和移位后的积分。 并针对所有shift值评估积分,从而生成卷积函数。

VIP内容

卷积作为CNN的基础部件之一,尽管常用,但存在两个主要缺陷:(1) 内容不可知;(2) 重度计算量。动态滤波器具有内容自适应特性,但同时进一步提升了计算量。深度(depth-wise)卷积是一种轻量型版本,但它往往会造成CNN性能下降,或者需要更大的通道数。

本文提出一种解耦动态滤波器(Decoupled Dynamic Filter),它可以同时解决上述缺陷。受启发于近期注意力的进展,DDF将深度动态滤波器解耦为空域与通道动态滤波器。该分解可以大大减少参数量,并将计算量限制在与深度卷积同等水平。同时,采用DDF替换分类网络中的标准卷积可以带来显著的性能提升。比如,ResNet50/101分别可以带来1.9%与1.3%的top1精度提升,且计算量近乎减半。在检测与联合上采样方面的实验同样证实了DDF上采样变种相比标准卷积的优异性。

本文所提DDF及其上采样变种DDF-Up具有以下几点优异属性:

Content-adaptive DDF提供了空间可变滤波器,这使得其具有内容自适应特性; Fast runtime DDF具有与深度卷积相近的计算量,因此它的推理速度要比标准卷积、动态滤波器更快; Smaller memory footprint DDF可以显著降低动态滤波器的内存占用,这使得我们可以采用DDF直接替换所有的标准卷积; Consistent performance improvements 采用DDF/DDF-Up替换标准卷积可以带来一致性的性能提升,同时在不同网络、不同任务上均取得了SOTA性能。

成为VIP会员查看完整内容
0
16

最新论文

Object recognition and viewpoint estimation lie at the heart of visual understanding. Recent works suggest that convolutional neural networks (CNNs) fail to generalize to out-of-distribution (OOD) category-viewpoint combinations, ie. combinations not seen during training. In this paper, we investigate when and how such OOD generalization may be possible by evaluating CNNs trained to classify both object category and 3D viewpoint on OOD combinations, and identifying the neural mechanisms that facilitate such OOD generalization. We show that increasing the number of in-distribution combinations (ie. data diversity) substantially improves generalization to OOD combinations, even with the same amount of training data. We compare learning category and viewpoint in separate and shared network architectures, and observe starkly different trends on in-distribution and OOD combinations, ie. while shared networks are helpful in-distribution, separate networks significantly outperform shared ones at OOD combinations. Finally, we demonstrate that such OOD generalization is facilitated by the neural mechanism of specialization, ie. the emergence of two types of neurons -- neurons selective to category and invariant to viewpoint, and vice versa.

0
0
下载
预览
父主题
Top