mxnet例子,图像识别

基本代码

主要完成的任务:

1. 加载相关包


2. 读取图片,并转换为2242243的RGB矩阵


3. 放入训练模型中


4. 得到预测结果

预先训练的模型包括1000个类,(y是1000个分类变量)分布于各种不同的动物

将原文件解压后运行示例代码:

library(mxnet)
library(imager)
# 载入模型
model = mx.model.load("Inception/Inception_BN", iteration=39)
# 载入mean image
mean.img = as.array(mx.nd.load("Inception/mean_224.nd")[["mean_img"]])

使用imager自带的鹦鹉图片来进行测试,数据读取,展示:

im <- load.image(system.file("extdata/parrots.png", package="imager"))
plot(im)

将输入图像转换为224*224的图像,作为模型输入,这部分不属于mxnet的范畴,可以看做是数据规范的过程

preproc.image <-function(im, mean.image) {
  # crop the image
  shape <- dim(im)
  short.edge <- min(shape[1:2])
  yy <- floor((shape[1] - short.edge) / 2) + 1
  yend <- yy + short.edge - 1
  xx <- floor((shape[2] - short.edge) / 2) + 1
  xend <- xx + short.edge - 1
  croped <- im[yy:yend, xx:xend,,]
  # resize to 224 x 224, needed by input of the model.
  resized <- resize(croped, 224, 224)
  # convert to array (x, y, channel)
  arr <- as.array(resized)
  dim(arr) = c(224, 224, 3)
  # substract the mean
  normed <- arr - mean.img
  # Reshape to format needed by mxnet (width, height, channel, num)
  dim(normed) <- c(224, 224, 3, 1)
  return(normed)
}

normed <- preproc.image(im, mean.img)

prob <- predict(model, X=normed)

max.idx <- max.col(t(prob))
synsets <- readLines("Inception/synset.txt")
print(paste0("Predicted Top-class: ", synsets[[max.idx]]))

## [1] "Predicted Top-class: n01818515 macaw"

showPic = function(input){
  cat <- load.image(system.file(input, package="imager"))
  plot(cat)
  normed <- preproc.image(cat, mean.img)
  prob <- predict(model, X=normed)
  max.idx <- max.col(t(prob))
  print(paste0("Predicted Top-class: ", synsets[[max.idx]]))
}

showPic("extdata/cat.jpg")

## [1] "Predicted Top-class: n02123597 Siamese cat, Siamese"

showPic("extdata/husky.jpg")

## [1] "Predicted Top-class: n02110185 Siberian husky"

showPic("extdata/labuladuo.jpg")

## [1] "Predicted Top-class: n02099712 Labrador retriever"