YOLO_tensorflow-master程式碼下載：

https://github.com/gliese581gg/YOLO_tensorflow

1.模型介紹

YOLO_tensorflow

(Version 0.2, Last updated :2016.02.16)

1.Introduction

This is tensorflow implementation of the YOLO:Real-Time Object Detection

It can only do predictions using pretrained YOLO_small & YOLO_tiny network for now.

I'm gonna support training later.

I extracted weight values from darknet's (.weight) files.

2.Install

(1) Download code

(2) Download YOLO weight file from

#YOLO_tensorflow

(Version 0.3, Last updated :2017.02.21)

###1.Introduction

This is tensorflow implementation of the YOLO:Real-Time Object Detection

It can only do predictions using pretrained YOLO_small & YOLO_tiny network for now.

I extracted weight values from darknet's (.weight) files.

My code does not support training. Use darknet for training.

###2.Install (1) Download code

(2) Download YOLO weight file from

(3) Put the 'YOLO_(version).ckpt' in the 'weight' folder of downloaded code

###3.Usage

(1) direct usage with default settings (display on console, show output image, no output file writing)

python YOLO_(small or tiny)_tf.py -fromfile (input image filename)

(2) direct usage with custom settings

python YOLO_(small or tiny)_tf.py argvs

where argvs are

-fromfile (input image filename) : input image file
-disp_console (0 or 1) : whether display results on terminal or not
-imshow (0 or 1) : whether display result image or not
-tofile_img (output image filename) : output image file
-tofile_txt (output txt filename) : output text file (contains class, x, y, w, h, probability)

(3) import on other scripts

import YOLO_(small or tiny)_tf
yolo = YOLO_(small or tiny)_tf.YOLO_TF()

yolo.disp_console = (True or False, default = True)
yolo.imshow = (True or False, default = True)
yolo.tofile_img = (output image filename)
yolo.tofile_txt = (output txt filename)
yolo.filewrite_img = (True or False, default = False)
yolo.filewrite_txt = (True of False, default = False)

yolo.detect_from_file(filename)
yolo.detect_from_cvmat(cvmat)

###4.Requirements

• Tensorflow
• Opencv2

###5.Copyright

According to the LICENSE file of the original code,

• Me and original author hold no liability for any damages
• Do not use this on commercial!

###6.Changelog 2016/02/15 : First upload!

2016/02/16 : Added YOLO_tiny, Fixed bug that ignores one of the boxes in grid when both boxes detected valid objects

2016/08/26 : Uploaded weight file converter! (darknet weight -> tensorflow ckpt)

(3) Put the 'YOLO_(version).ckpt' in the 'weight' folder of downloaded code

3.Usage

(1) direct usage with default settings (display on console, show output image, no output file writing)

python YOLO_(small or tiny)_tf.py -fromfile (input image filename)

(2) direct usage with custom settings

python YOLO_(small or tiny)_tf.py argvs

where argvs are

-fromfile (input image filename) : input image file
-disp_console (0 or 1) : whether display results on terminal or not
-imshow (0 or 1) : whether display result image or not
-tofile_img (output image filename) : output image file
-tofile_txt (output txt filename) : output text file (contains class, x, y, w, h, probability)

(3) import on other scripts

import YOLO_(small or tiny)_tf
yolo = YOLO_(small or tiny)_tf.YOLO_TF()

yolo.disp_console = (True or False, default = True)
yolo.imshow = (True or False, default = True)
yolo.tofile_img = (output image filename)
yolo.tofile_txt = (output txt filename)
yolo.filewrite_img = (True or False, default = False)
yolo.filewrite_txt = (True of False, default = False)

yolo.detect_from_file(filename)
yolo.detect_from_cvmat(cvmat)

4.Requirements

• Tensorflow
• Opencv2

5.Copyright

According to the LICENSE file of the original code,

• Me and original author hold no liability for any damages
• Do not use this on commercial!

6.Changelog

2016/02/15 : First upload!

2016/02/16 : Added YOLO_tiny, Fixed bug that ignores one of the boxes in grid when both boxes detected valid objects

2016/08/26 : Uploaded weight file converter! (darknet weight -> tensorflow ckpt)

2.模型使用

我使用YOLO時

出現

cv2.imread('./test/person.jpg')

解決辦法 在程式碼最前面加

import cv2

下面是我調研yolo的所有程式碼。

#encoding:utf-8
import cv2
import YOLO_tiny_tf


yolo = YOLO_tiny_tf.YOLO_TF()

yolo.disp_console = True
yolo.imshow = True
yolo.tofile_img = './test/ttt.jpg'
yolo.tofile_txt = './test/ttt.txt'
yolo.filewrite_img = True
yolo.filewrite_txt = True
filename = './test/person.jpg'
# 讀入影象
#im = cv2.imread('./test/person.jpg')yolo.detect_from_file(filename)
#yolo.detect_from_cvmat(im)

執行結果

3.模型儲存與執行

（1）.將YOLO_ting_tf.py中的輸入新增名字input，程式碼如下：

def build_networks(self):
   if self.disp_console : print "Building YOLO_tiny graph..."
self.x = tf.placeholder('float32',[None,448,448,3],name="input")

（2）.執行save_graph模型和權重一起儲存

import os
import cv2

import tensorflow as tf
import numpy as np

from tensorflow.python.framework import test_util
import freeze_graph

from YOLO_tiny_tf import YOLO_TF


def save_graph(sess,output_path,checkpoint,checkpoint_state_name,input_graph_name,output_graph_name):

    checkpoint_prefix = os.path.join(output_path,checkpoint)
    saver = tf.train.Saver(tf.all_variables())
    saver.save(sess, checkpoint_prefix, global_step=0,latest_filename=checkpoint_state_name)
    tf.train.write_graph(sess.graph.as_graph_def(),output_path,
                           input_graph_name)

    # We save out the graph to disk, and then call the const conversion
    # routine.
input_graph_path = os.path.join(output_path, input_graph_name)
    input_saver_def_path = ""
input_binary = False
input_checkpoint_path = checkpoint_prefix + "-0"
output_node_names = "19_fc"
restore_op_name = "save/restore_all"
filename_tensor_name = "save/Const:0"
output_graph_path = os.path.join(output_path, output_graph_name)
    clear_devices = False
freeze_graph.freeze_graph(input_graph_path, input_saver_def_path,
                              input_binary, input_checkpoint_path,
                              output_node_names, restore_op_name,
                              filename_tensor_name, output_graph_path,clear_devices, "")


yolo=YOLO_TF()
#with open("weights/small_model.pb","wb") as f:
 #   f.write(yolo.sess.graph.as_graph_def().SerializeToString())
save_graph(yolo.sess,"/home/acer/pbMake/yolo","saved_checkpoint","checkpoint_state","yoloting_input_graph.pb","yoloting_output_graph.pb")

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import cv2

import tensorflow as tf
import numpy as np



def iou(box1, box2):
    tb = min(box1[0] + 0.5 * box1[2], box2[0] + 0.5 * box2[2]) - max(box1[0] - 0.5 * box1[2], box2[0] - 0.5 * box2[2])
    lr = min(box1[1] + 0.5 * box1[3], box2[1] + 0.5 * box2[3]) - max(box1[1] - 0.5 * box1[3], box2[1] - 0.5 * box2[3])
    if tb < 0 or lr < 0:
        intersection = 0
else:
        intersection = tb * lr
    return intersection / (box1[2] * box1[3] + box2[2] * box2[3] - intersection)

def interpret_output(output):
    alpha = 0.1
threshold = 0.2
iou_threshold = 0.5
num_class = 20
num_box = 2
grid_size = 7
classes = ["aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat", "chair", "cow", "diningtable",
               "dog", "horse", "motorbike", "person", "pottedplant", "sheep", "sofa", "train", "tvmonitor"]

    w_img = 640
h_img = 480
probs = np.zeros((7, 7, 2, 20))
    class_probs = np.reshape(output[0:980], (7, 7, 20))
    scales = np.reshape(output[980:1078], (7, 7, 2))
    boxes = np.reshape(output[1078:], (7, 7, 2, 4))
    offset = np.transpose(np.reshape(np.array([np.arange(7)] * 14), (2, 7, 7)), (1, 2, 0))

    boxes[:, :, :, 0] += offset
    boxes[:, :, :, 1] += np.transpose(offset, (1, 0, 2))
    boxes[:, :, :, 0:2] = boxes[:, :, :, 0:2] / 7.0
boxes[:, :, :, 2] = np.multiply(boxes[:, :, :, 2], boxes[:, :, :, 2])
    boxes[:, :, :, 3] = np.multiply(boxes[:, :, :, 3], boxes[:, :, :, 3])

    boxes[:, :, :, 0] *= w_img
    boxes[:, :, :, 1] *= h_img
    boxes[:, :, :, 2] *= w_img
    boxes[:, :, :, 3] *= h_img

    for i in range(2):
        for j in range(20):
            probs[:, :, i, j] = np.multiply(class_probs[:, :, j], scales[:, :, i])

    filter_mat_probs = np.array(probs >= threshold, dtype='bool')
    filter_mat_boxes = np.nonzero(filter_mat_probs)
    boxes_filtered = boxes[filter_mat_boxes[0], filter_mat_boxes[1], filter_mat_boxes[2]]
    probs_filtered = probs[filter_mat_probs]
    classes_num_filtered = np.argmax(filter_mat_probs, axis=3)[
        filter_mat_boxes[0], filter_mat_boxes[1], filter_mat_boxes[2]]

    argsort = np.array(np.argsort(probs_filtered))[::-1]
    boxes_filtered = boxes_filtered[argsort]
    probs_filtered = probs_filtered[argsort]
    classes_num_filtered = classes_num_filtered[argsort]

    for i in range(len(boxes_filtered)):
        if probs_filtered[i] == 0: continue
        for j in range(i + 1, len(boxes_filtered)):
            if iou(boxes_filtered[i], boxes_filtered[j]) > iou_threshold:
                probs_filtered[j] = 0.0
filter_iou = np.array(probs_filtered > 0.0, dtype='bool')
    boxes_filtered = boxes_filtered[filter_iou]
    probs_filtered = probs_filtered[filter_iou]
    classes_num_filtered = classes_num_filtered[filter_iou]

    result = []
    for i in range(len(boxes_filtered)):
        result.append(
            [classes[classes_num_filtered[i]], boxes_filtered[i][0], boxes_filtered[i][1], boxes_filtered[i][2],
             boxes_filtered[i][3], probs_filtered[i]])

    return result


def show_results(img, results):
    filewrite_img = False
filewrite_txt = True
img_cp = img.copy()
    if filewrite_txt:
        ftxt = open('./test/xsss.txt', 'w')
    for i in range(len(results)):
        x = int(results[i][1])
        y = int(results[i][2])
        w = int(results[i][3]) // 2
h = int(results[i][4]) // 2
cv2.rectangle(img_cp, (x - w, y - h), (x + w, y + h), (0, 255, 0), 2)
        cv2.rectangle(img_cp, (x - w, y - h - 20), (x + w, y - h), (125, 125, 125), -1)
        cv2.putText(img_cp, results[i][0] + ' : %.2f' % results[i][5], (x - w + 5, y - h - 7),
        cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 0), 1)

        ftxt.write(results[i][0] + ',' + str(x) + ',' + str(y) + ',' + str(w) + ',' + str(h) + ',' + str(
                results[i][5]) + '\n')


        cv2.imwrite('./test/xlsld.jpg', img_cp)


# produces the expected result.
with tf.Graph().as_default():
    output_graph_def = tf.GraphDef()
    output_graph_path = '/home/acer/pbMake/yolo/yoloting_output_graph.pb'
x = tf.placeholder('float32', [None, 448, 448, 3])

    with open(output_graph_path, "rb") as f:
        output_graph_def.ParseFromString(f.read())
        _ = tf.import_graph_def(output_graph_def, name="")


    with tf.Session() as sess:
        tf.initialize_all_variables().run()
        input_x = sess.graph.get_tensor_by_name("input:0")
        print(input_x)
        output = sess.graph.get_tensor_by_name("19_fc:0")
        print(output)

        filename = './test/person.jpg'
img = cv2.imread(filename)

        h_img, w_img, _ = img.shape
        img_resized = cv2.resize(img, (448, 448))
        img_RGB = cv2.cvtColor(img_resized, cv2.COLOR_BGR2RGB)
        img_resized_np = np.asarray(img_RGB)
        inputs = np.zeros((1, 448, 448, 3), dtype='float32')
        inputs[0] = (img_resized_np / 255.0) * 2.0 - 1.0
#input_node = sess.graph.get_operation_by_name("input")
in_dict = {input_x: inputs}
        net_output = sess.run(output, {input_x: inputs})
        print("net_output", net_output)


        #net_output = sess.run(output_node, feed_dict=in_dict)
result = interpret_output(net_output[0])
        show_results(img, result)

執行結果，與直接執行模型的結果一樣。

yolo（7） 

目錄(?)[+]

本系列文章會持續更新，主要會分以下幾個部分：  １、darknet下的yolo原始碼解讀  ２、將yolo移植到mxnet下  ３、模型壓縮與加速  白天需要工作，只有晚上時間寫，所以可能更新速度有點慢，還有就是該系列博文不一定會嚴格按照以上三點的順序來寫，也可能移植到caffe下，在caffe下進行壓縮和加速。

一、訓練

我用的是VOC2007的資料集，下載指令如下：

$curl -O http://pjreddie.com/media/files/VOCtrainval_06-Nov-2007.tar
$curl -O http://pjreddie.com/media/files/VOCtest_06-Nov-2007.tar
$tar xf VOCtrainval_06-Nov-2007.tar
$tar xf VOCtest_06-Nov-2007.tar

• 1
• 2
• 3
• 4

• 1
• 2
• 3
• 4

執行以下程式碼，將.xml 檔案轉換成.txt 檔案，以備YOLO訓練時資料解析：

import xml.etree.ElementTree as ET
import pickle
import os
from os import listdir, getcwd
from os.path import join

sets=[('2007', 'train'), ('2007', 'val')]

classes = ["aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat", "chair", "cow", "diningtable", "dog", "horse", "motorbike", "person", "pottedplant", "sheep", "sofa", "train", "tvmonitor"]


def convert(size, box):
    dw = 1./size[0]
    dh = 1./size[1]
    x = (box[0] + box[1])/2.0
    y = (box[2] + box[3])/2.0
    w = box[1] - box[0]
    h = box[3] - box[2]
    x = x*dw
    w = w*dw
    y = y*dh
    h = h*dh
    return (x,y,w,h)

def convert_annotation(year, image_id):
    in_file = open('VOCdevkit/VOC%s/Annotations/%s.xml'%(year, image_id))
    out_file = open('VOCdevkit/VOC%s/labels/%s.txt'%(year, image_id), 'w')
    tree=ET.parse(in_file)
    root = tree.getroot()
    size = root.find('size')
    w = int(size.find('width').text)
    h = int(size.find('height').text)

    for obj in root.iter('object'):
        difficult = obj.find('difficult').text
        cls = obj.find('name').text
        if cls not in classes or int(difficult) == 1:
            continue
        cls_id = classes.index(cls)
        xmlbox = obj.find('bndbox')
        b = (float(xmlbox.find('xmin').text), float(xmlbox.find('xmax').text), float(xmlbox.find('ymin').text), float(xmlbox.find('ymax').text))
        bb = convert((w,h), b)
        out_file.write(str(cls_id) + " " + " ".join([str(a) for a in bb]) + '\n')

wd = getcwd()

for year, image_set in sets:
    if not os.path.exists('VOCdevkit/VOC%s/labels/'%(year)):
        os.makedirs('VOCdevkit/VOC%s/labels/'%(year))
    image_ids = open('VOCdevkit/VOC%s/ImageSets/Main/%s.txt'%(year, image_set)).read().strip().split()
    list_file = open('%s_%s.txt'%(year, image_set), 'w')
    for image_id in image_ids:
        list_file.write('%s/VOCdevkit/VOC%s/JPEGImages/%s.jpg\n'%(wd, year, image_id))
        convert_annotation(year, image_id)
    list_file.close()

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16
• 17
• 18
• 19
• 20
• 21
• 22
• 23
• 24
• 25
• 26
• 27
• 28
• 29
• 30
• 31
• 32
• 33
• 34
• 35
• 36
• 37
• 38
• 39
• 40
• 41
• 42
• 43
• 44
• 45
• 46
• 47
• 48
• 49
• 50
• 51
• 52
• 53
• 54
• 55

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16
• 17
• 18
• 19
• 20
• 21
• 22
• 23
• 24
• 25
• 26
• 27
• 28
• 29
• 30
• 31
• 32
• 33
• 34
• 35
• 36
• 37
• 38
• 39
• 40
• 41
• 42
• 43
• 44
• 45
• 46
• 47
• 48
• 49
• 50
• 51
• 52
• 53
• 54
• 55

在yolo.c中找到以下程式碼，並修改：

//指向你剛才生成的train.txt檔案
char *train_images = "/data/voc/train.txt";
//新建個資料夾，然後指向這裡就可以，訓練生成的模型檔案會儲存在這裡
char *backup_directory = "/home/pjreddie/backup/";

• 1
• 2
• 3
• 4

• 1
• 2
• 3
• 4

執行以下指令開始訓練：

./darknet yolo train cfg/yolo.train.cfg extraction.conv.weights

• 1

• 1

二、原始碼解讀

1.首先我們看一下訓練的資料流，從main函式開始看，該函式在darknet.c 檔案中：

//darknet.c
int main(int argc, char **argv)
{
    //test_resize("data/bad.jpg");
    //test_box();
    //test_convolutional_layer();
    if(argc < 2){
        fprintf(stderr, "usage: %s <function>\n", argv[0]);
        return 0;
    }
    gpu_index = find_int_arg(argc, argv, "-i", 0);
    if(find_arg(argc, argv, "-nogpu")) {
        gpu_index = -1;
    }

#ifndef GPU
    gpu_index = -1;
#else
    if(gpu_index >= 0){
        cuda_set_device(gpu_index);
    }
#endif

    if (0 == strcmp(argv[1], "average")){
        average(argc, argv);
    } else if (0 == strcmp(argv[1], "yolo")){
        //第一個引數是yolo，所以跳轉到run_yolo函式
        run_yolo(argc, argv);
    } else {
        fprintf(stderr, "Not an option: %s\n", argv[1]);
    }
    return 0;
}

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16
• 17
• 18
• 19
• 20
• 21
• 22
• 23
• 24
• 25
• 26
• 27
• 28
• 29
• 30
• 31
• 32
• 33

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16
• 17
• 18
• 19
• 20
• 21
• 22
• 23
• 24
• 25
• 26
• 27
• 28
• 29
• 30
• 31
• 32
• 33

//yolo.c
void run_yolo(int argc, char **argv)
{
    char *prefix = find_char_arg(argc, argv, "-prefix", 0);
    float thresh = find_float_arg(argc, argv, "-thresh", .2);
    int cam_index = find_int_arg(argc, argv, "-c", 0);
    int frame_skip = find_int_arg(argc, argv, "-s", 0);
    if(argc < 4){
        fprintf(stderr, "usage: %s %s [train/test/valid] [cfg] [weights (optional)]\n", argv[0], argv[1]);
        return;
    }

    char *cfg = argv[3];
    char *weights = (argc > 4) ? argv[4] : 0;
    char *filename = (argc > 5) ? argv[5]: 0;
    if(0==strcmp(argv[2], "test")) test_yolo(cfg, weights, filename, thresh);
    //第二個引數是train，所以跳轉到了train_yolo函式
    else if(0==strcmp(argv[2], "train")) train_yolo(cfg, weights);
    voc_names, 20, frame_skip, prefix);
}


void train_yolo(char *cfgfile, char *weightfile)
{
    char *train_images = "/data/voc/train.txt";
    char *backup_directory = "/home/pjreddie/backup/";
    /*srand函式是隨機數發生器的初始化函式。
    srand和rand()配合使用產生偽隨機數序列。rand函式在產生隨機數前，需要系統提供的生成偽隨機數序列的
    種子，rand根據這個種子的值產生一系列隨機數。如果系統提供的種子沒有變化，每次呼叫rand函式生成的偽
    隨機數序列都是一樣的。*/
    srand(time(0));
    /*第三個引數是：`cfg/yolo.train.cfg`，`basecfg()`這個函式把`cfg/yolo.train.cfg`變成了
    `yolo0train.cfg`，然後用base指標指向`yolo0train.cfg`*/
    char *base = basecfg(cfgfile);
    //列印"yolo"字樣
    printf("%s\n", base);
    float avg_loss = -1;
    //解析網路構架，下面會仔細分析該函式
    network net = parse_network_cfg(cfgfile);
    //載入預訓練引數，下面會仔細分析該函式
    if(wei