update 126_130 Face&Pose&YOLO

Author: JimmyHHua

README.md

@@ -163,4 +163,9 @@ code_121 | [DNN- Read the information of each layer of the model](python/code_12
 code_122 | [DNN- Realize image classification](python/code_122)   | ✔️
 code_123 | [DNN- Model runs to set the target device and compute the background](python/code_123)   | ✔️
 code_124 | [DNN- SSD Single Image Detection](python/code_124)   | ✔️
-code_125 | [DNN- SSD Real-time Video Detection](python/code_125)   | ✔️
+code_125 | [DNN- SSD Real-time Video Detection](python/code_125)   | ✔️
+code_126 | [DNN- Face Detection based on Residual Network](python/code_126)   | ✔️
+code_127 | [DNN- Video Face Detection based on Residual Network](python/code_127)   | ✔️
+code_128 | [DNN- Call the Detection Model of Tensorflow](python/code_128)   | ✔️
+code_129 | [DNN- Call the Openpose Implementation Attitude Assessment](python/code_129)   | ✔️
+code_130 | [DNN- Call YOLO Object Detection Network](python/code_130)   | ✔️

README_CN.md

@@ -162,4 +162,9 @@ code_121 | [DNN- 读取模型各层信息](python/code_121)   | ✔️
 code_122 | [DNN- DNN实现图像分类](python/code_122)   | ✔️
 code_123 | [DNN- 模型运行设置目标设备与计算后台](python/code_123)   | ✔️
 code_124 | [DNN- SSD单张图片检测](python/code_124)   | ✔️
-code_125 | [DNN- SSD实时视频检测](python/code_125)   | ✔️
+code_125 | [DNN- SSD实时视频检测](python/code_125)   | ✔️
+code_126 | [DNN- 基于残差网络的人脸检测](python/code_126)   | ✔️
+code_127 | [DNN- 基于残差网络的视频人脸检测](python/code_127)   | ✔️
+code_128 | [DNN- 调用tensorflow的检测模型](python/code_128)   | ✔️
+code_129 | [DNN- 调用openpose模型实现姿态评估](python/code_129)   | ✔️
+code_130 | [DNN- 调用YOLO对象检测网络](python/code_130)   | ✔️

python/code_126/README.md

@@ -0,0 +1,51 @@
+## OpenCV 基于残差网络的人脸检测
+
+### 概述
+
+✔️ ️OpenCV在DNN模块中提供了基于残差SSD网络训练的人脸检测模型，该模型分别提供了tensorflow版本，caffe版本，torch版本模型文件。
+
+✔️ 其中tensorflow版本的模型做了更加进一步的压缩优化，大小只有2MB左右，非常适合移植到移动端使用，实现人脸检测功能，而caffe版本的是fp16的浮点数模型，精准度更好。
+
+✔️ 对比传统人脸检测，同样一张图像，在OpenCV HAAR与LBP级联检测器中必须通过不断调整参数才可以检测出全部人脸，而通过使用该模型，基本在Python语言中基于OpenCV后台的推断，在25毫秒均可以检测出结果，网络支持输入size大小为300x300。
+
+### 代码
+```python
+import cv2
+
+model_bin = "../model/face_detector/opencv_face_detector_uint8.pb";
+config_text = "../model/face_detector/opencv_face_detector.pbtxt";
+
+# load tensorflow model
+net = cv2.dnn.readNetFromTensorflow(model_bin, config=config_text)
+image = cv2.imread("people.jpg")
+h = image.shape[0]
+w = image.shape[1]
+
+# 人脸检测
+blobImage = cv2.dnn.blobFromImage(image, 1.0, (300, 300), (104.0, 177.0, 123.0), False, False);
+net.setInput(blobImage)
+Out = net.forward()
+
+t, _ = net.getPerfProfile()
+label = 'Inference time: %.2f ms' % (t * 1000.0 / cv2.getTickFrequency())
+cv2.putText(image, label, (0, 20), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)
+
+# 绘制检测矩形
+for detection in Out[0,0,:,:]:
+    score = float(detection[2])
+    objIndex = int(detection[1])
+    if score > 0.5:
+        left = detection[3]*w
+        top = detection[4]*h
+        right = detection[5]*w
+        bottom = detection[6]*h
+
+        # 绘制
+        cv2.rectangle(image, (int(left), int(top)), (int(right), int(bottom)), (255, 0, 0), thickness=2)
+        cv2.putText(image, "score:%.2f"%score, (int(left), int(top)-1), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1)
+
+cv2.imshow('demo', image)
+```
+>输出
+
+<img src="./result.jpg">

python/code_126/opencv_126.py

@@ -0,0 +1,37 @@
+import cv2
+
+model_bin = "../model/face_detector/opencv_face_detector_uint8.pb";
+config_text = "../model/face_detector/opencv_face_detector.pbtxt";
+
+# load tensorflow model
+net = cv2.dnn.readNetFromTensorflow(model_bin, config=config_text)
+image = cv2.imread("face.jpg")
+h = image.shape[0]
+w = image.shape[1]
+
+# 人脸检测
+blobImage = cv2.dnn.blobFromImage(image, 1.0, (300, 300), (104.0, 177.0, 123.0), False, False);
+net.setInput(blobImage)
+Out = net.forward()
+
+t, _ = net.getPerfProfile()
+label = 'Inference time: %.2f ms' % (t * 1000.0 / cv2.getTickFrequency())
+cv2.putText(image, label, (0, 20), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)
+
+# 绘制检测矩形
+for detection in Out[0,0,:,:]:
+    score = float(detection[2])
+    objIndex = int(detection[1])
+    if score > 0.5:
+        left = detection[3]*w
+        top = detection[4]*h
+        right = detection[5]*w
+        bottom = detection[6]*h
+
+        # 绘制
+        cv2.rectangle(image, (int(left), int(top)), (int(right), int(bottom)), (255, 0, 0), thickness=2)
+        cv2.putText(image, "score:%.2f"%score, (int(left), int(top)-1), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1)
+
+cv2.imshow('face-detection-demo', image)
+cv2.waitKey(0)
+cv2.destroyAllWindows()

python/code_126/people.jpg

@@ -0,0 +1,64 @@
+## OpenCV 基于残差网络的视频人脸检测
+
+### 概述
+
+✔️ OpenCV在DNN模块中提供了基于残差SSD网络训练的人脸检测模型，它支持单精度的fp16的检测，准确度更好的Caffe模型加载与使用.
+
+
+### 代码
+
+✔️ 这里实现了一个基于Caffe Model的视频实时人脸监测模型，基于Python, 在CPU运行，可以达到fps16以上。
+
+```python
+import cv2
+
+model_bin = "../model/face_detector/res10_300x300_ssd_iter_140000_fp16.caffemodel";
+config_text = "../model/face_detector/deploy.prototxt";
+
+# load caffe model
+net = cv2.dnn.readNetFromCaffe(config_text, model_bin)
+
+# set back-end
+net.setPreferableBackend(cv2.dnn.DNN_BACKEND_OPENCV)
+net.setPreferableTarget(cv2.dnn.DNN_TARGET_CPU)
+
+cap = cv2.VideoCapture(0)
+while True:
+    ret, image = cap.read()
+    image = cv2.flip(image, 1)
+    if ret is False:
+        break
+    # 人脸检测
+    h, w = image.shape[:2]
+    blobImage = cv2.dnn.blobFromImage(image, 1.0, (300, 300), (104.0, 177.0, 123.0), False, False);
+    net.setInput(blobImage)
+    cvOut = net.forward()
+
+    # Put efficiency information.
+    t, _ = net.getPerfProfile()
+    fps = 1000 / (t * 1000.0 / cv2.getTickFrequency())
+    label = 'FPS: %.2f' % fps
+    cv2.putText(image, label, (0, 15), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0))
+
+    # 绘制检测矩形
+    for detection in cvOut[0,0,:,:]:
+        score = float(detection[2])
+        objIndex = int(detection[1])
+        if score > 0.5:
+            left = detection[3]*w
+            top = detection[4]*h
+            right = detection[5]*w
+            bottom = detection[6]*h
+
+            # 绘制
+            cv2.rectangle(image, (int(left), int(top)), (int(right), int(bottom)), (255, 0, 0), thickness=2)
+            cv2.putText(image, "score:%.2f"%score, (int(left), int(top)), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1)
+    cv2.imshow('face-detection-demo', image)
+    c = cv2.waitKey(2)
+    if c == 27:
+        break
+cv2.waitKey(0)
+cv2.destroyAllWindows()
+```
+
+<img src="./song.gif">

python/code_126/result.jpg

@@ -0,0 +1,49 @@
+import cv2
+
+model_bin = "../model/face_detector/res10_300x300_ssd_iter_140000_fp16.caffemodel";
+config_text = "../model/face_detector/deploy.prototxt";
+
+# load caffe model
+net = cv2.dnn.readNetFromCaffe(config_text, model_bin)
+
+# set back-end
+net.setPreferableBackend(cv2.dnn.DNN_BACKEND_OPENCV)
+net.setPreferableTarget(cv2.dnn.DNN_TARGET_CPU)
+
+cap = cv2.VideoCapture('song.mp4')
+while True:
+    ret, image = cap.read()
+    image = cv2.flip(image, 1)
+    if ret is False:
+        break
+    # 人脸检测
+    h, w = image.shape[:2]
+    blobImage = cv2.dnn.blobFromImage(image, 1.0, (300, 300), (104.0, 177.0, 123.0), False, False);
+    net.setInput(blobImage)
+    cvOut = net.forward()
+
+    # Put efficiency information.
+    t, _ = net.getPerfProfile()
+    fps = 1000 / (t * 1000.0 / cv2.getTickFrequency())
+    label = 'FPS: %.2f' % fps
+    cv2.putText(image, label, (0, 15), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0))
+
+    # 绘制检测矩形
+    for detection in cvOut[0,0,:,:]:
+        score = float(detection[2])
+        objIndex = int(detection[1])
+        if score > 0.5:
+            left = detection[3]*w
+            top = detection[4]*h
+            right = detection[5]*w
+            bottom = detection[6]*h
+
+            # 绘制
+            cv2.rectangle(image, (int(left), int(top)), (int(right), int(bottom)), (255, 0, 0), thickness=2)
+            cv2.putText(image, "score:%.2f"%score, (int(left), int(top)), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1)
+    cv2.imshow('face-detection-demo', image)
+    c = cv2.waitKey(2)
+    if c == 27:
+        break
+cv2.waitKey(0)
+cv2.destroyAllWindows()