ImageNet Classification with Deep Convolutional Neural Networks -- 解读

论文解读

先稍微翻译下

数据集

ImageNet 有各种分辨率的图片，我们需要固定大小的。所以，向下取样到 256 * 256。除此之外没有别的预处理。
ImageNet consists of variable-resolution images, while our system requires a constant input dimen- sionality. Therefore, we down-sampled the images to a fixed resolution of 256 × 256. Given a rectangular image, we first rescaled the image such that the shorter side was of length 256, and then cropped out the central 256×256 patch from the resulting image. We did not pre-process the images in any other way, except for subtracting the mean activity over the training set from each pixel. So we trained our network on the (centered) raw RGB values of the pixels.

网络结构

• 采用RELU激活函数

• Local Response Normalization
  

• Overlapping Pooling

减少overfitting

• data augmentation
  如何增加数据呢？
  首先，随机选取 224 * 224 patches和它们的 their horizontal reflections

第二个方法包括 改变RGB 通道的密度，
The second form of data augmentation consists of altering the intensities of the RGB channels in training images. Specifically, we perform PCA on the set of RGB pixel values throughout the ImageNet training set.

To each training image, we add multiples of the found principal components，with magnitudes proportional to the corresponding eigenvalues times a random variable drawn from a Gaussian with mean zero and standard deviation 0.1.

Dropout

代码