Food for Thought — Paper Tuesday

栏目: IT技术 · 发布时间: 4年前

内容简介:Every Tuesday I highlight an interesting paper that I came across in research or work. I hope that my review can help you get the juiciest part of the paper under 2 minutes!It is widely accepted that CNNs learn images by extracting shape features like curv

Food for Thought — Paper Tuesday

Use Shape-biased Data Improves Accuracy and Robustness

Every Tuesday I highlight an interesting paper that I came across in research or work. I hope that my review can help you get the juiciest part of the paper under 2 minutes!

Basic Ideas

It is widely accepted that CNNs learn images by extracting shape features like curves and edges. However, a group of researchers from University of Tuebingen and Edinburgh challenged this belief in their ICLR 2019 paper ImageNet-Trained CNNs Are Biased Towards Texture; Increasing Shape Bias Improves Accuracy and Robustness .

From the paper

Here’s the link: https://openreview.net/pdf?id=Bygh9j09KX

By cleverly crafting several experiments, the researchers demonstrated CNNs are more biased toward image textures than people would expect. From that, they further found that shape-enhanced dataset can serve as an effective data augmentation method that improves model accuracy and robustness.

The researchers argued that CNNs are heavily biased toward local features, perhaps due to the small perception field of convolution filters. Their argument is supported by CNN’s surprisingly low performance on texture-free images demonstrated in the image below

From the paper

As demonstrated in the figure, all major-league architectures like AlexNet, GoogleNet, and VGG16 experience significant performance drops when texture information is removed (silhouette and edge). Meanwhile, CNNs yield high confidence even when shape information is removed as long as texture is present (texture).

Results

In order to further test their hypothesis, the researchers generated a new dataset called Stylized-ImageNet (SIN), whose images’ local texture features are replaced by uninformative random features.

From the paper

If CNNs are biased toward local texture features, we would expect CNNs trained on the original ImageNet dataset to perform poorly on SIN. This is indeed the case, as demonstrated in the following table

From the paper

Okay, now we are convinced that CNNs are biased toward local textures. But how can we use this inforamtion to our advantage? The researchers demonstrated that models jointly trained on SIN and IN is more robust to image distortion (noise, cropping, filtering…) and achieve outstanding accuracies in image classification and object detection.

From the paper
From the paper

Some Thoughts

For a long time I was convinced that CNNs are capable of image classification because of its powerful edge detectors. This paper opened a new door for us — there are some many handwaving explaination and understanding of neural networks. There are still a lot of theoretical work ahead to understand even one of the simplest forms of neural networks!


以上就是本文的全部内容,希望本文的内容对大家的学习或者工作能带来一定的帮助,也希望大家多多支持 码农网

查看所有标签

猜你喜欢:

本站部分资源来源于网络,本站转载出于传递更多信息之目的,版权归原作者或者来源机构所有,如转载稿涉及版权问题,请联系我们

Google模式

Google模式

Eric Schmidt、Jonathan Rosenberg / 李芳齡 / 天下雜誌出版社 / 2014-11-7 / TWD 420.00

上市即登紐約時報暢銷書、Amazon科技經營排行榜TOP1 未上市即售出美、英、德、日、荷等12國版權 Google創辦人Larry Page專文推薦 第一本由Google領導團隊人首度公開的企業內部運作與思維 Google董事會執行主席艾力克.施密特獨家揭露 Google從崛起到稱霸超過10年的管理與工作秘笈, 以及如何吸引21世紀最搶手的人才-智慧創做者(S......一起来看看 《Google模式》 这本书的介绍吧!

CSS 压缩/解压工具
CSS 压缩/解压工具

在线压缩/解压 CSS 代码

Markdown 在线编辑器
Markdown 在线编辑器

Markdown 在线编辑器

html转js在线工具
html转js在线工具

html转js在线工具