dlib下训练自己的物体检测器--手的检测

之前我们在Linux上安装了dlib(http://www.cnblogs.com/take-fetter/p/8318602.html),也成功的完成了之前的人脸检测程序,

今天我们来一起学习怎样使用dlib创建属于自己的简单的物体识别器(这里以手的检测为例，特别感谢https://handmap.github.io/dlib-classifier-for-object-detection/)

• imglab的介绍与安装

imglab是dlib提供的个工具,位于github dlib开源项目的tools目录下.imglab是一个简单的图形工具，用对象边界来标注图像盒子和可选的零件位置。 一般来说，你可以在需要时使用它

以训练物体检测器（例如脸部检测器），因为它允许你轻松创建所需的训练数据集。

(源码位于https://github.com/davisking/dlib/tree/master/tools/imglab 如果有兴趣使用的话建议先下载整个dlib项目并安装dlib后再对本工具进行编译)

　　　编译依次使用

 cd dlib/tools/imglab
    mkdir build
    cd build
    cmake ..
    cmake --build . --config Release

　　 不建议使用readme.txt中关于sudo make install的命令,因为我使用之后出现了无法显示图像的错误

• 训练自己的手检测器（关于手的图片的dataset可以参考Hand Images Databases - https://www.mutah.edu.jo/biometrix/hand-images-databases.html提供的数据集

　　　　　　　　　　　　 或http://www.robots.ox.ac.uk/~vgg/data/hands/的相关数据集）

　　使用cmake后的build文件目录下（windows则位于release目录中）完成如下操作

　　使用

./imglab -c mydataset.xml 图片目录

　　创建mydataset.xml完成创建mydataset.xml 和image_metadata_stylesheet.xsl的样式表

　　使用

./imglab mydataset.xml

　　会打开一个窗口，这里就需要对每张图片进行位置的框选，在Next Label中输入框选信息，并对每张图片进行框选（按住shift并鼠标左键点击拖动画框）

　　在将对图片全标注后，在files选项中点击save，我们便可以关闭窗口，此时打开mydataset.xml可以看到其中包含了图片信息，如图

　　

之后将mydataset.xml 和image_metadata_stylesheet.xsl放入图片目录中，运行如下代码进行训练（可能会出现图片目录出错的情况，这里需要对mydataset.xml中的图片位置进行确认）

代码改自dlib的python_examples，如果要自己尝试，建议先认真看下github中的代码（https://github.com/davisking/dlib/blob/master/python_examples/train_object_detector.py）

运行程序需使用scikit-image使用pip install scikit-image 安装

import os
import sys
import glob

import dlib
from skimage import io


# In this example we are going to train a face detector based on the small
# faces dataset in the examples/faces directory.  This means you need to supply
# the path to this faces folder as a command line argument so we will know
# where it is.
if len(sys.argv) != 2:
    print(
        "Give the path to the examples/faces directory as the argument to this "
        "program. For example, if you are in the python_examples folder then "
        "execute this program by running:\n"
        "    ./train_object_detector.py ../examples/faces")
    exit()
faces_folder = sys.argv[1]


# Now let's do the training.  The train_simple_object_detector() function has a
# bunch of options, all of which come with reasonable default values.  The next
# few lines goes over some of these options.
options = dlib.simple_object_detector_training_options()
# Since faces are left/right symmetric we can tell the trainer to train a
# symmetric detector.  This helps it get the most value out of the training
# data.
options.add_left_right_image_flips = True
# The trainer is a kind of support vector machine and therefore has the usual
# SVM C parameter.  In general, a bigger C encourages it to fit the training
# data better but might lead to overfitting.  You must find the best C value
# empirically by checking how well the trained detector works on a test set of
# images you haven't trained on.  Don't just leave the value set at 5.  Try a
# few different C values and see what works best for your data.
options.C = 5
# Tell the code how many CPU cores your computer has for the fastest training.
options.num_threads = 4
options.be_verbose = True


training_xml_path = os.path.join(faces_folder, "palm-landmarks.xml")
testing_xml_path = os.path.join(faces_folder, "testing.xml")
# This function does the actual training.  It will save the final detector to
# detector.svm.  The input is an XML file that lists the images in the training
# dataset and also contains the positions of the face boxes.  To create your
# own XML files you can use the imglab tool which can be found in the
# tools/imglab folder.  It is a simple graphical tool for labeling objects in
# images with boxes.  To see how to use it read the tools/imglab/README.txt
# file.  But for this example, we just use the training.xml file included with
# dlib.
dlib.train_simple_object_detector(training_xml_path, "detector.svm", options)

接下来就是等待训练完成（当然在这里说下，数据集不宜过大，会导致内存不足而OS自动杀死线程/进程的情况），options中的参数很多需要自行根据情况调节的

训练完成后会生成detector.svm文件，使用如下程序进行一个简单的测试：

import imutils
import dlib
import cv2
import time

detector = dlib.simple_object_detector("detector_from_author.svm")

image = cv2.imread('test0.jpg')
image = imutils.resize(image, width=500)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

rects = detector(gray, 1)
#win_det = dlib.image_window()
#win_det.set_image(detector)

#win = dlib.image_window()

for (k, d) in enumerate(rects):
    print("Detection {}: Left: {} Top: {} Right: {} Bottom: {}".format(
        k, d.left(), d.top(), d.right(), d.bottom()))
    cv2.rectangle(image, (d.left(), d.top()), (d.right(), d.bottom()), (0, 255, 0), 2)

#win.add_overlay(rects)
cv2.imshow("Output", image)
cv2.waitKey(0)

运行结果

可以看到完成了手的检测。

 

 

后记：

1. 训练时间很长，希望能耐心等待
2. 再次特别感谢Nathan Glover以及他的教程https://handmap.github.io/dlib-classifier-for-object-detection/
3. 如果要制作精度很高的检测器，并不建议使用本方法，因为我们最终生成的svm文件相比于dlib作者的人脸识别检测器而言相差甚远。
4. 我认为dlib提供的imglab功能很少，不适用于大规模的需要高精度的识别情况（不过人脸识别还是很不错的）
5. 对于需要高精度高准确率的物体识别，使用Tensorflow Object Detection API应该更为合适（https://github.com/tensorflow/models/tree/master/research/object_detection）