解读 Data Augmentation using Random Image Cropping and Patching for Deep CNNs
2018-12-25 21:32
1191 查看
论文提出了一种新的数据增强的方法:利用随机图像裁剪和拼接的方法random image cropping and patching (RICAP)
文章链接:https://arxiv.org/pdf/1811.09030.pdf
在人工智能日益流行的趋向下,数据显得尤为重要。 4000 既能增加样本多样性,也能防止过拟合。现有的数据增强方法有:翻转flipping,缩放resizing,平移变换shifting等
本文提出RICAP可看下图直观理解:
RICAP 是一种新的数据增强的方法,可以运用在深层次的卷积神经网络中。另外,RICAP也提到关于标签平滑。
RICAP主要有三个步骤:
- 从训练集中随机选取四张图片
- 分别裁剪各张图片
- 将裁剪得到的图片拼接成为一张新的图片
文中给出的python代码如下:
涉及numpy和 PyTorch模块
相关文章推荐
- 论文解读| Very Deep Convolutional Networks for Large-Scale Image Recognition
- (转)Image Segmentation with Tensorflow using CNNs and Conditional Random Fields
- Using a Callout for Deep Inspection of Stream Data
- 笔记:Deep multi patch aggregation network for image style, aesthetics and quality estimation
- 深度学习论文随记(二)---VGGNet模型解读-2014年(Very Deep Convolutional Networks for Large-Scale Image Recognition)
- [翻译]Effect of Sensor Fusion for Recognition of Emotional States Using Voice, Face Image and Thermal Image of Face
- 译Combining Markov Random Fields and Convolutional Neural Networks for Image Synthesis
- 阅读笔记—Robust Image Sentiment Analysis using Progressively Trained and Domain Transferred Deep Network
- 论文笔记:Improving Deep Visual Representation for Person Re-identification by Global and Local Image-language Association
- produce a gradient file using surface elevation data, and plot it using grdimage with topographic "shade"
- 解读Deep Residual Learning for Image recognition
- Deep Residual Learning for Image Recognition(ResNet)残差网络解读
- Sensor fusion and input representation for time series classification using deep nets
- Learning Deep Structured Semantic Models for Web Search using Clickthrough Data笔记
- Exploring Models and Data for Image Question Answering
- 论文PCANet: A Simple Deep Learning Baseline for Image Classification?的matlab源码解读(四)
- Bit-Scalable Deep Hashing with Regularized Similarity Learning for Image Retrieval and Person Re-ide
- Along with all the above benefits, you cannot overlook the space efficiency and performance gains in using DataFrames and Dataset APIs for two reasons.
- [论文笔记]Learning Deep Structured Semantic Models for Web Search using Clickthrough Data
- 论文PCANet: A Simple Deep Learning Baseline for Image Classification?的matlab源码解读(一)