首先得到了https://blog.csdn.net/gybheroin/article/details/72581318系列博客的帮助。表示感激。
关于安装caffe已在之前的博客介绍，自用可行，https://www.cnblogs.com/MY0213/p/9225310.html
1.数据源
首先使用的数据集为人脸数据集,可在百度云自行下载：
链接：https://pan.baidu.com/s/156DiOuB46wKrM0cEaAgfMw 密码：1ap0
将train.zip解压可得数据源，label文件是val.txt和train.txt。
2.将图片数据做成lmdb数据源
SET GLOG_logtostderr=1 SET RESIZE_HEIGHT=227 SET RESIZE_WIDTH=227 "convert_imageset" --resize_height=227 --resize_width=227 --shuffle "train/" "train.txt" "mtraindb" "convert_imageset" --resize_height=227 --resize_width=227 --shuffle "val/" "val.txt" "mvaldb" pause
详见face_lmdb.bat,将数据做成同等大小的数据。
3. 得到图像均值
SET GLOG_logtostderr=1 "compute_image_mean" "mtraindb" "train_mean.binaryproto" pause
详见mean_face.bat
训练时先做减均值的操作，可能对训练效果有好处
这里可以用固定的图片均值，是多少可以直接百度谷歌，这一步也可以不做，唐宇迪大神说影响不大。
4. 图像训练
SET GLOG_logostderr=1 caffe train --solver=solver.prototxt pause
详见train.bat
net: "train.prototxt" test_iter: 100 test_interval: 1000 # lr for fine-tuning should be lower than when starting from scratch base_lr: 0.001 lr_policy: "step" gamma: 0.1 # stepsize should also be lower, as we're closer to being done stepsize: 1000 display: 50 max_iter: 10000 momentum: 0.9 weight_decay: 0.0005 snapshot: 1000 snapshot_prefix: "model" # uncomment the following to default to CPU mode solving # solver_mode: CPU
详见solver.prototxt
关于solver.prototxt的内涵可查看
https://blog.csdn.net/qq_27923041/article/details/55211808
############################# DATA Layer #############################
name: "face_train_val"
layer {
top: "data"
top: "label"
name: "data"
type: "Data"
data_param {
source: "mtraindb"
backend:LMDB
batch_size: 64
}
transform_param {
mean_file: "train_mean.binaryproto"
mirror: true
}
include: { phase: TRAIN }
}
layer {
top: "data"
top: "label"
name: "data"
type: "Data"
data_param {
source: "mvaldb"
backend:LMDB
batch_size: 64
}
transform_param {
mean_file: "train_mean.binaryproto"
mirror: true
}
include: {
phase: TEST
}
}
layer {
name: "conv1"
type: "Convolution"
bottom: "data"
top: "conv1"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 96
kernel_size: 11
stride: 4
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "relu1"
type: "ReLU"
bottom: "conv1"
top: "conv1"
}
layer {
name: "norm1"
type: "LRN"
bottom: "conv1"
top: "norm1"
lrn_param {
local_size: 5
alpha: 0.0001
beta: 0.75
}
}
layer {
name: "pool1"
type: "Pooling"
bottom: "norm1"
top: "pool1"
pooling_param {
pool: MAX
kernel_size: 3
stride: 2
}
}
layer {
name: "conv2"
type: "Convolution"
bottom: "pool1"
top: "conv2"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 256
pad: 2
kernel_size: 5
group: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu2"
type: "ReLU"
bottom: "conv2"
top: "conv2"
}
layer {
name: "norm2"
type: "LRN"
bottom: "conv2"
top: "norm2"
lrn_param {
local_size: 5
alpha: 0.0001
beta: 0.75
}
}
layer {
name: "pool2"
type: "Pooling"
bottom: "norm2"
top: "pool2"
pooling_param {
pool: MAX
kernel_size: 3
stride: 2
}
}
layer {
name: "conv3"
type: "Convolution"
bottom: "pool2"
top: "conv3"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 384
pad: 1
kernel_size: 3
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "relu3"
type: "ReLU"
bottom: "conv3"
top: "conv3"
}
layer {
name: "conv4"
type: "Convolution"
bottom: "conv3"
top: "conv4"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 384
pad: 1
kernel_size: 3
group: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu4"
type: "ReLU"
bottom: "conv4"
top: "conv4"
}
layer {
name: "conv5"
type: "Convolution"
bottom: "conv4"
top: "conv5"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 256
pad: 1
kernel_size: 3
group: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu5"
type: "ReLU"
bottom: "conv5"
top: "conv5"
}
layer {
name: "pool5"
type: "Pooling"
bottom: "conv5"
top: "pool5"
pooling_param {
pool: MAX
kernel_size: 3
stride: 2
}
}
layer {
name: "fc6"
type: "InnerProduct"
bottom: "pool5"
top: "fc6"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 4096
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu6"
type: "ReLU"
bottom: "fc6"
top: "fc6"
}
layer {
name: "drop6"
type: "Dropout"
bottom: "fc6"
top: "fc6"
dropout_param {
dropout_ratio: 0.5
}
}
layer {
name: "fc7"
type: "InnerProduct"
bottom: "fc6"
top: "fc7"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 4096
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu7"
type: "ReLU"
bottom: "fc7"
top: "fc7"
}
layer {
name: "drop7"
type: "Dropout"
bottom: "fc7"
top: "fc7"
dropout_param {
dropout_ratio: 0.5
}
}
layer {
name: "fc8-expr"
type: "InnerProduct"
bottom: "fc7"
top: "fc8-expr"
param {
lr_mult: 10
decay_mult: 1
}
param {
lr_mult: 20
decay_mult: 0
}
inner_product_param {
num_output: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "accuracy"
type: "Accuracy"
bottom: "fc8-expr"
bottom: "label"
top: "accuracy"
include {
phase: TEST
}
}
layer {
name: "loss"
type: "SoftmaxWithLoss"
bottom: "fc8-expr"
bottom: "label"
top: "loss"
}
详见train.prototxt，也就是将alexnet中最后的1000变为2就可以了。
这个过程需要5天左右（我用的cpu）,可以直接用已有模型alexnet_iter_50000_full_conv.caffemodel
5. 测试
可用run_face_detect_batch.py测试人脸检测效果。
6. 总结
这个网络测试时特别慢，用的是slipping window的方法。下面的文章再介绍快速一点的faster rcnn 及FPN。
slipping window中用了Casting a Classifier into a Fully Convolutional Network 的方法。这一方法在其他网络中也可用。
关于rcnn的演进，可见https://www.cnblogs.com/MY0213/p/9460562.html
欢迎批评指正。