版權所有，翻版必究。
運行環境：WIN10，pycharm，相應的CUDA，CUDNN，tensorflow1.15.0，tensorflow-gpu-1.14.0，Anaconda3
（很多問題呢都出現在環境的配置上面，所以可以的話開個虛擬環境進行訓練處理）
使用python或者C++調用接口來處理，其中python可以看我之前的版本。調用.h5文件即可。本文中主要介紹一個嗶哩嗶哩大佬使用.pb和.pbtxt來完成調用。主要講解的是C++和python調用的.pb文件和.pbtxt文件的生成問題。後續調用，看代碼即可。（稍安毋昭，看下文）
鏈接在此：https://www.bilibili.com/video/BV1jJ411u7A5
大家可以去看這個人的視頻，很詳細。我這裏主要是用於記錄和自己學習，也歡迎大家一起學習。
https://www.learnopencv.com/deep-learning-based-object-detection-and-instance-segmentation-using-mask-r-cnn-in-opencv-python-c/ （這裏放的是另外一個大佬對於此類問題的分析）
下載文件：
鏈接：https://pan.baidu.com/s/1tI-_S-GJSW3WcMKsdHTBkQ
提取碼：xk8p
打開後爲：
第一步配置虛擬環境：
打開anoconda下的：
輸入命令：（配置tensorflow） conda create --name tensorflow python=3.7回車即可。我都是因爲已經裝好了，所以不需要再裝虛擬環境。
你們應該會出現這種樣子：輸入y即可。
然後會出現提示：
激活該環境的命令就在其中，使用conda activate tensorflow即可激活環境。然後你們的命令行會變成這樣：
表明已經激活了。
虛擬環境配置好了之後就需要配置安裝包。看你的電腦屬於那種，選擇其中一種進行配置。
//CPU
conda install tensorflow=1.14.0
//GPU
conda install tensorflow-gpu=1.14.0
配置opencv opencv ipython cython pillow：conda install opencv ipython cython pillow
第二步：環境配置
#環境配置
在你安裝的（Anacondal路徑下）.\Anaconda3\envs\tensorflow\Lib\site-packages
新建tensorflow.pth ,添加
.\models-master\research
.\models-master\research\slim
這兩個文件的目錄。（需要看你的壓縮文件放在哪裏了！！！）
#cocoapi
cd cocoapi/PythonAPI
python setup.py build_ext install
（需要配置PythonAPI目錄下的文件，裏面放置的是一些依賴庫和文件）
第三步：轉換數據
python create_tf_record.py --images_dir=./datasets/train/image --annotations_json_dir=./datasets/train/json --label_map_path=./datasets/label.pbtxt --output_path=./output/train.record
python create_tf_record.py --images_dir=./datasets/val/image --annotations_json_dir=./datasets/val/json --label_map_path=./datasets/label.pbtxt --output_path=./output/val.record
第四步：訓練
#訓練，修改mask_rcnn_inception_v2_coco.config裏面標註有Need Modified
python ./research/object_detection/model_main.py --model_dir=./output --pipeline_config_path=./finetune_model/mask_rcnn_inception_v2_coco.config
第五步：轉換模型（生成.pb文件和.pbtxt文件）
#轉換模型,按需要改路徑名字和文件名字
python ./research/object_detection/export_inference_graph.py --input_type image_tensor --pipeline_config_path ./finetune_model/mask_rcnn_inception_v2_coco.config --trained_checkpoint_prefix ./output/model.ckpt-1000 --output_directory ./output
python tf_text_graph_mask_rcnn.py --input ./output/frozen_inference_graph.pb --output ./output/mask_rcnn.pbtxt --config ./finetune_model/mask_rcnn_inception_v2_coco.config
python mask_rcnn_predict.py
重頭戲在他的視頻中使用C++的opencv來調用這塊，給了我很大的啓發。代碼在他的這個目錄下：
我會添加一些自己的理解在其中，覺得有問題的歡迎說明討論！（不喜勿噴謝謝！）
主程序：
#include "stdafx.h"
#include <fstream>
#include <sstream>
#include <iostream>
#include <string.h>
#include <opencv2/dnn.hpp>
#include <opencv2/imgproc.hpp>
#include <opencv2/highgui.hpp>
using namespace cv;
using namespace dnn;
using namespace std;
String textGraph = "mask_rcnn.pbtxt";
String modelWeights = "frozen_inference_graph.pb";
string classesFile = "mscoco_labels.names";
string colorsFile = "colors.txt";
float confThreshold = 0.7;
float maskThreshold = 0.3;
int ImgWidth = 224;
int ImgHight = 224;
vector<string> classes;
vector<Scalar> colors;
Mat frame, blob, m_DstMat;
float m_fWidthScale;
float m_fHeighScale;
void Predect(string path);
void LoadLabelAndColor();
void drawBox(Mat& frame, int classId, float conf, Rect box, Mat& objectMask);
void postprocess(Mat& frame, const vector<Mat>& outs);
int main(int argc, char** argv)
{
LoadLabelAndColor();
Predect("1.jpg");
return 0;
}
void Predect(string path) {
// Load the network
Net net = readNetFromTensorflow(modelWeights, textGraph);
net.setPreferableBackend(DNN_BACKEND_OPENCV);
net.setPreferableTarget(DNN_TARGET_CPU);
frame = imread(path);
m_DstMat = frame.clone();
resize(frame, frame, Size(ImgWidth, ImgHight));
m_fWidthScale = m_DstMat.cols*1.0 / frame.cols;
m_fHeighScale = m_DstMat.rows*1.0 / frame.rows;
// Stop the program if reached end of video
if (frame.empty()) {
return ;
}
// Create a 4D blob from a frame.
blobFromImage(frame, blob, 1.0, Size(frame.cols, frame.rows), Scalar(), true, false);
//blobFromImage(frame, blob);
net.setInput(blob);
std::vector<String> outNames(2);
outNames[0] = "detection_out_final";
outNames[1] = "detection_masks";
vector<Mat> outs;
net.forward(outs, outNames);
postprocess(frame, outs);
vector<double> layersTimes;
double freq = getTickFrequency() / 1000;
double t = net.getPerfProfile(layersTimes) / freq;
string label = format("Inference time for a frame : %0.0f ms", t);
putText(m_DstMat, label, Point(0, 15), FONT_HERSHEY_SIMPLEX, 0.5, Scalar(0, 0, 0));
imshow("Result", m_DstMat);
waitKey(0);
}
此處主程序沒有寫python版本，但是要弄python版本的話也就是改成對應的語句，難度不大。想嘗試的可以試試。我這篇博客主要用於自己學習記錄。
//此處就是利用程序預測出來的邊框點進行圖像的繪製。當然也會設置閾值對結果進行一定程度的把控！
// For each frame, extract the bounding box and mask for each detected object
void postprocess(Mat& frame, const vector<Mat>& outs)
{
Mat outDetections = outs[0];
Mat outMasks = outs[1];
// Output size of masks is NxCxHxW where
// N - number of detected boxes
// C - number of classes (excluding background)
// HxW - segmentation shape
const int numDetections = outDetections.size[2];
const int numClasses = outMasks.size[1];
cout<< "numClasses: "<<numDetections << "numClasses: "<<numClasses <<endl;
outDetections = outDetections.reshape(1, outDetections.total() / 7);
for (int i = 0; i < numDetections; ++i)
{
float score = outDetections.at<float>(i, 2);
if (score > confThreshold)
{
// Extract the bounding box
int classId = static_cast<int>(outDetections.at<float>(i, 1));
int left = static_cast<int>(frame.cols * outDetections.at<float>(i, 3));
int top = static_cast<int>(frame.rows * outDetections.at<float>(i, 4));
int right = static_cast<int>(frame.cols * outDetections.at<float>(i, 5));
int bottom = static_cast<int>(frame.rows * outDetections.at<float>(i, 6));
left = max(0, min(left, frame.cols - 1));
top = max(0, min(top, frame.rows - 1));
right = max(0, min(right, frame.cols - 1));
bottom = max(0, min(bottom, frame.rows - 1));
Rect box = Rect(left, top, right - left + 1, bottom - top + 1);
/************************************************/
box.x = round(box.x*m_fWidthScale);
box.y = round(box.y*m_fHeighScale);
box.width = round(box.width*m_fWidthScale);
box.height = round(box.height*m_fHeighScale);
/************************************************/
// Extract the mask for the object
Mat objectMask(outMasks.size[2], outMasks.size[3], CV_32F, outMasks.ptr<float>(i, classId));
// Draw bounding box, colorize and show the mask on the image
drawBox(m_DstMat, classId, score, box, objectMask);
}
}
}
python版本：
# For each frame, extract the bounding box and mask for each detected object
def postprocess(boxes, masks):
# Output size of masks is NxCxHxW where
# N - number of detected boxes
# C - number of classes (excluding background)
# HxW - segmentation shape
numClasses = masks.shape[1]
numDetections = boxes.shape[2]
frameH = frame.shape[0]
frameW = frame.shape[1]
for i in range(numDetections):
box = boxes[0, 0, i]
mask = masks[i]
score = box[2]
if score > confThreshold:
classId = int(box[1])
# Extract the bounding box
left = int(frameW * box[3])
top = int(frameH * box[4])
right = int(frameW * box[5])
bottom = int(frameH * box[6])
left = max(0, min(left, frameW - 1))
top = max(0, min(top, frameH - 1))
right = max(0, min(right, frameW - 1))
bottom = max(0, min(bottom, frameH - 1))
# Extract the mask for the object
classMask = mask[classId]
# Draw bounding box, colorize and show the mask on the image
drawBox(frame, classId, score, left, top, right, bottom, classMask)
//這一塊的主要目的是在輸入圖像上，利用預測的掩模圖以及邊框點信息進行圖像的繪製。
// Draw the predicted bounding box, colorize and show the mask on the image
void drawBox(Mat& frame, int classId, float conf, Rect box, Mat& objectMask)
{
//Draw a rectangle displaying the bounding box
rectangle(frame, Point(box.x, box.y), Point(box.x + box.width, box.y + box.height), Scalar(255, 178, 50), 3);
//Get the label for the class name and its confidence
string label = format("%.2f", conf);
if (!classes.empty())
{
CV_Assert(classId < (int)classes.size());
label = classes[classId] + ":" + label;
}
//Display the label at the top of the bounding box
int baseLine;
Size labelSize = getTextSize(label, FONT_HERSHEY_SIMPLEX, 0.5, 1, &baseLine);
box.y = max(box.y, labelSize.height);
rectangle(frame, Point(box.x, box.y - round(1.5*labelSize.height)), Point(box.x + round(1.5*labelSize.width), box.y + baseLine), Scalar(255, 255, 255), FILLED);
putText(frame, label, Point(box.x, box.y), FONT_HERSHEY_SIMPLEX, 0.75, Scalar(0, 0, 0), 1);
Scalar color = colors[classId%colors.size()];
// Resize the mask, threshold, color and apply it on the image
resize(objectMask, objectMask, Size(box.width, box.height));
Mat mask = (objectMask > maskThreshold);
Mat coloredRoi = (0.3 * color + 0.7 * frame(box));
coloredRoi.convertTo(coloredRoi, CV_8UC3);
// Draw the contours on the image
vector<Mat> contours;
Mat hierarchy;
mask.convertTo(mask, CV_8U);
findContours(mask, contours, hierarchy, RETR_CCOMP, CHAIN_APPROX_SIMPLE);
drawContours(coloredRoi, contours, -1, color, 5, LINE_8, hierarchy, 100);
coloredRoi.copyTo(frame(box), mask);
}
python版本：
# Draw the predicted bounding box, colorize and show the mask on the image
def drawBox(frame, classId, conf, left, top, right, bottom, classMask):
# Draw a bounding box.
cv.rectangle(frame, (left, top), (right, bottom), (255, 178, 50), 3)
# Print a label of class.
label = '%.2f' % conf
if classes:
assert(classId < len(classes))
label = '%s:%s' % (classes[classId], label)
# Display the label at the top of the bounding box
labelSize, baseLine = cv.getTextSize(label, cv.FONT_HERSHEY_SIMPLEX, 0.5, 1)
top = max(top, labelSize[1])
cv.rectangle(frame, (left, top - round(1.5*labelSize[1])), (left + round(1.5*labelSize[0]), top + baseLine), (255, 255, 255), cv.FILLED)
cv.putText(frame, label, (left, top), cv.FONT_HERSHEY_SIMPLEX, 0.75, (0,0,0), 1)
# Resize the mask, threshold, color and apply it on the image
classMask = cv.resize(classMask, (right - left + 1, bottom - top + 1))
mask = (classMask > maskThreshold)
roi = frame[top:bottom+1, left:right+1][mask]
color = colors[classId%len(colors)]
# Comment the above line and uncomment the two lines below to generate different instance colors
#colorIndex = random.randint(0, len(colors)-1)
#color = colors[colorIndex]
frame[top:bottom+1, left:right+1][mask] = ([0.3*color[0], 0.3*color[1], 0.3*color[2]] + 0.7 * roi).astype(np.uint8)
# Draw the contours on the image
mask = mask.astype(np.uint8)
im2, contours, hierarchy = cv.findContours(mask,cv.RETR_TREE,cv.CHAIN_APPROX_SIMPLE)
cv.drawContours(frame[top:bottom+1, left:right+1], contours, -1, color, 3, cv.LINE_8, hierarchy, 100)
//這一塊的本質並不複雜，就是將文件中的內容讀取出來，讓其符合C++使用的規則。
//第一步讀取文件
//第二步讀取文件中的內容
//第三步放置到你需要存儲的地方。比如此段代碼中放置的位置爲classes和colors
//這也是爲了後面畫圖的時候，用顏色區分來進行提前準備的
void LoadLabelAndColor()
{
ifstream ifs(classesFile.c_str());
string line;
while (getline(ifs, line)) classes.push_back(line);
ifstream colorFptr(colorsFile.c_str());
while (getline(colorFptr, line))
{
char* pEnd;
double r, g, b;
r = strtod(line.c_str(), &pEnd);
g = strtod(pEnd, NULL);
b = strtod(pEnd, NULL);
Scalar color = Scalar(r, g, b, 255.0);
colors.push_back(Scalar(r, g, b, 255.0));
}
}
對應的python代碼爲：
# Load names of classes
classesFile = "mscoco_labels.names";
classes = None
with open(classesFile, 'rt') as f:
classes = f.read().rstrip('\n').split('\n')
# Load the colors
colorsFile = "colors.txt";
with open(colorsFile, 'rt') as f:
colorsStr = f.read().rstrip('\n').split('\n')
colors = []
for i in range(len(colorsStr)):
rgb = colorsStr[i].split(' ')
color = np.array([float(rgb[0]), float(rgb[1]), float(rgb[2])])
colors.append(color)
stdafx.cpp中存放的東西如下所示：
// stdafx.cpp : 只包括標準包含文件的源文件
// MaskRCNNCPP.pch 將作爲預編譯頭
// stdafx.obj 將包含預編譯類型信息
#include "stdafx.h"
// TODO: 在 STDAFX.H 中引用任何所需的附加頭文件，
//而不是在此文件中引用
附一張結果圖：
參考了很多大佬的文章，盡情期待我都下一篇文章。該文章主要是說明，如何在C#中使用opencvsharp使用該模型！！！（(●'◡'●)這纔是博主的本來目的）。這邊主要也是給予大家一點好多文章。day day up！