Caffe 实现多标签分类

最近在用Ｃａｆｆｅ做验证码识别时，发现没有用Tensorflow简单(tensorflow中可以用one-hot, 参考我的另一篇blog:  http://blog.csdn.net/sushiqian/article/details/78305340 )，需要修改Ｃａｆｆｅ的源码，做完后，觉得也不复杂。

1. 首先，就是修改源码了，

    本文介绍的方法是修改ImageDataLayer，修改下面的三个文件：

          $CAFFE_ROOT/src/caffe/proto/caffe.proto

          $CAFFE_ROOT/include/caffe/layers/image_data_layer.hpp

          $CAFFE_ROOT/src/caffe/layers/image_data_layer.cpp

          (1) 修改　caffe.proto

                在文件　caffe.proto 里的　message ImageDataParameter {  } 里添加新的一项，用于表示标签的维度

　　　　optional int32 label_dim = 13 [default = 1];

　　


　　(2) 修改　image_data_layer.hpp

　　　把原先成员  lines_  的定义修改为

　　　vector<std::pair<std::string, vector<int> > > lines_;



　　(３) 修改　image_data_layer.cpp

          修改函数　DataLayerSetUp()

          新的代码如下：

int label_dim = this->layer_param_.image_data_param().label_dim();
string filename;
while (infile >> filename){
vector<int> label(label_dim, 0);
for(int i = 0; i < label_dim; i++) {
infile >> label[i];
}
lines_.push_back(std::make_pair(filename, label));
}

vector<int> label_shape(2);
label_shape[0] = batch_size;
label_shape[1] = label_dim;

　　修改函数　load_batch()

int label_dim = image_data_param.label_dim();

for(int i = 0;i < label_dim;++i){
prefetch_label[item_id * label_dim + i] = lines_[lines_id_].second[i];
}

   (4) 用　make　编译

　　$CAFFE_ROOT/build$ make

        $CAFFE_ROOT/build$ make install

2. 编写用于训练的网络结构

　可以在caffe examples 里的某个ｎｅｔ上修改一个，　例如，本文中所讲例子是在mnist中LeNet（lenet_train_test.prototxt）基础上修改出一个net

     captcha_train_test.prototxt

name: "captcha"
layer {
name: "Input"
type: "ImageData"
top: "data"
top: "label"
include {
phase: TRAIN
}
transform_param {
scale: 0.00390625
}
image_data_param {
source: "/home/mark/data/train.txt"
root_folder: "/home/mark/data/"
new_height: 60
new_width: 160
is_color: true
batch_size: 50
shuffle: true
label_dim: 4
}
}

layer {
name: "Input"
type: "ImageData"
top: "data"
top: "label"
include {
phase: TEST
}
transform_param {
scale: 0.00390625
}
image_data_param {
source: "/home/mark/data/test.txt"
root_folder: "/home/mark/data/"
new_height: 60
new_width: 160
is_color: true
batch_size: 20
shuffle: true
label_dim: 4
}
}

layer {
name: "slice"
type: "Slice"
bottom: "label"
top: "label_1"
top: "label_2"
top: "label_3"
top: "label_4"
slice_param {
axis: 1
slice_point:1
slice_point:2
slice_point:3
}
}

layer {
name: "conv1"
type: "Convolution"
bottom: "data"
top: "conv1"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
convolution_param {
num_output: 20
kernel_size: 5
stride: 1
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}
layer {
name: "pool1"
type: "Pooling"
bottom: "conv1"
top: "pool1"
pooling_param {
pool: MAX
kernel_size: 2
stride: 2
}
}
layer {
name: "conv2"
type: "Convolution"
bottom: "pool1"
top: "conv2"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
convolution_param {
num_output: 50
kernel_size: 5
stride: 1
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}
layer {
name: "pool2"
type: "Pooling"
bottom: "conv2"
top: "pool2"
pooling_param {
pool: MAX
kernel_size: 2
stride: 2
}
}
layer {
name: "ip1"
type: "InnerProduct"
bottom: "pool2"
top: "ip1"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 500
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}
layer {
name: "relu1"
type: "ReLU"
bottom: "ip1"
top: "ip1"
}

layer {
name: "ip2"
type: "InnerProduct"
bottom: "ip1"
top: "ip2"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 100
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}

layer {
name: "ip3_1"
type: "InnerProduct"
bottom: "ip2"
top: "ip3_1"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 10
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}

layer {
name: "ip3_2"
type: "InnerProduct"
bottom: "ip2"
top: "ip3_2"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 10
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}

layer {
name: "ip3_3"
type: "InnerProduct"
bottom: "ip2"
top: "ip3_3"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 10
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}

layer {
name: "ip3_4"
type: "InnerProduct"
bottom: "ip2"
top: "ip3_4"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 10
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}

layer {
name: "accuracy1"
type: "Accuracy"
bottom: "ip3_1"
bottom: "label_1"
top: "accuracy1"
include {
phase: TEST
}
}
layer {
name: "loss1"
type: "SoftmaxWithLoss"
bottom: "ip3_1"
bottom: "label_1"
top: "loss1"
}

layer {
name: "accuracy2"
type: "Accuracy"
bottom: "ip3_2"
bottom: "label_2"
top: "accuracy2"
include {
phase: TEST
}
}
layer {
name: "loss2"
type: "SoftmaxWithLoss"
bottom: "ip3_2"
bottom: "label_2"
top: "loss2"
}

layer {
name: "accuracy3"
type: "Accuracy"
bottom: "ip3_3"
bottom: "label_3"
top: "accuracy3"
include {
phase: TEST
}
}
layer {
name: "loss3"
type: "SoftmaxWithLoss"
bottom: "ip3_3"
bottom: "label_3"
top: "loss3"
}

layer {
name: "accuracy4"
type: "Accuracy"
bottom: "ip3_4"
bottom: "label_4"
top: "accuracy4"
include {
phase:
4000
TEST
}
}
layer {
name: "loss4"
type: "SoftmaxWithLoss"
bottom: "ip3_4"
bottom: "label_4"
top: "loss4"
}

/home/mark/data/train.txt　里面存放的是训练所用图片的路径和标签，



该网络的结构图如下所示：



solver 文件

captcha_solver.prototxt

# The train/test net protocol buffer definition
net: "examples/captcha/captcha_train_test.prototxt"
# test_iter specifies how many forward passes the test should carry out.
# covering the full 9,800 testing images.
test_iter: 200
# Carry out testing every 200 training iterations.
test_interval: 200
# The base learning rate, momentum and the weight decay of the network.
base_lr: 0.001
momentum: 0.9
weight_decay: 0.0005
# The learning rate policy
lr_policy: "inv"
gamma: 0.0001
power: 0.75
# Display every 100 iterations
display: 100
# The maximum number of iterations
max_iter: 10000
# snapshot intermediate results
snapshot: 5000
snapshot_prefix: "examples/captcha/captcha"
# solver mode: CPU or GPU
solver_mode: GPU

在CAFFE_ROOT 下执行下面的命令开始训练

./build/tools/caffe train --solver=examples/captcha/captcha_solver.prototxt



Note: 训练过程中如果发现 loss 不下降，可以把学习率（base_lr）调小一点试试

3. 模型训练成功后，进行测试

　首先，需要一个deploy.prototxt文件，在captcha_train_test.prototxt的基础上修改，修改后保存为　captcha_deploy.prototxt  内容如下

name: "captcha"

input: "data"
input_dim: 1   # batchsize
input_dim: 3   # number of channels - rgb
input_dim: 60  # height
input_dim: 160 # width

layer {
name: "conv1"
type: "Convolution"
bottom: "data"
top: "conv1"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
convolution_param {
num_output: 20
kernel_size: 5
stride: 1
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}
layer {
name: "pool1"
type: "Pooling"
bottom: "conv1"
top: "pool1"
pooling_param {
pool: MAX
kernel_size: 2
stride: 2
}
}
layer {
name: "conv2"
type: "Convolution"
bottom: "pool1"
top: "conv2"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
convolution_param {
num_output: 50
kernel_size: 5
stride: 1
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}
layer {
name: "pool2"
type: "Pooling"
bottom: "conv2"
top: "pool2"
pooling_param {
pool: MAX
kernel_size: 2
stride: 2
}
}
layer {
name: "ip1"
type: "InnerProduct"
bottom: "pool2"
top: "ip1"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 500
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}
layer {
name: "relu1"
type: "ReLU"
bottom: "ip1"
top: "ip1"
}

layer {
name: "ip2"
type: "InnerProduct"
bottom: "ip1"
top: "ip2"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 100
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}

layer {
name: "ip3_1"
type: "InnerProduct"
bottom: "ip2"
top: "ip3_1"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 10
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}

layer {
name: "ip3_2"
type: "InnerProduct"
bottom: "ip2"
top: "ip3_2"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 10
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}

layer {
name: "ip3_3"
type: "InnerProduct"
bottom: "ip2"
top: "ip3_3"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 10
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}

layer {
name: "ip3_4"
type: "InnerProduct"
bottom: "ip2"
top: "ip3_4"
param {
lr_mult: 1
}
param {
lr_mult: 2
}
inner_product_param {
num_output: 10
weight_filler {
type: "xavier"
}
bias_filler {
type: "constant"
}
}
}

layer {
name: "prob1"
type: "Softmax"
bottom: "ip3_1"
top: "prob1"
}
layer {
name: "prob2"
type: "Softmax"
bottom: "ip3_2"
top: "prob2"
}
layer {
name: "prob3"
type: "Softmax"
bottom: "ip3_3"
top: "prob3"
}
layer {
name: "prob4"
type: "Softmax"
bottom: "ip3_4"
top: "prob4"
}

编写测试代码：

import numpy as np
import os
import sys
os.environ['GLOG_minloglevel'] = '3'
import caffe

CAFFE_ROOT = '/home/mark/caffe'
deploy_file_name = 'captcha_deploy.prototxt'
model_file_name  = 'captcha_iter_10000.caffemodel'

IMAGE_HEIGHT = 60
IMAGE_WIDTH = 160
IMAGE_CHANNEL = 3

def classify(imageFileName):
deploy_file = CAFFE_ROOT + '/examples/captcha/' + deploy_file_name
model_file  = CAFFE_ROOT + '/examples/captcha/' + model_file_name
#初始化caffe
net = caffe.Net(deploy_file, model_file, caffe.TEST)

#数据预处理
transformer = caffe.io.Transformer({'data': net.blobs['data'].data.shape})
transformer.set_transpose('data', (2, 0, 1))#pycaffe读取的图片文件格式为H×W×C，需转化为C×H×W

#pycaffe将图片存储为[0, 1], 如果模型输入用的是0~255的原始格式,需要做如下转换
#transformer.set_raw_scale('data', 255)

transformer.set_channel_swap('data', (2, 1, 0))#caffe中图片是BGR格式，而原始格式是RGB，所以要转化

# 将输入图片格式转化为合适格式（与deploy文件相同）
net.blobs['data'].reshape(1, IMAGE_CHANNEL, IMAGE_HEIGHT, IMAGE_WIDTH)

#读取图片
#参数color: True(default)是彩色图，False是灰度图
img = caffe.io.load_image(imageFileName, color=True)

#数据输入、预处理
net.blobs['data'].data[...] = transformer.preprocess('data', img)

#前向迭代，即分类
out = net.forward()

#求出每个标签概率最大值的下标
result = []
predict1 = out['prob1'][0].argmax()
result.append(predict1)

predict2 = out['prob2'][0].argmax()
result.append(predict2)

predict3 = out['prob3'][0].argmax()
result.append(predict3)

predict4 = out['prob4'][0].argmax()
result.append(predict4)

return result

if __name__ == '__main__':

imgList = sys.argv[1:]
for captcha in imgList:
predict = classify(captcha)
print "captcha:", captcha, "  predict:", predict

执行上面的脚本，进行测试



测试用到的图片如下