OpenPose 基于OpenCV DNN 的多人姿态估计

原文: OpenPose 基于OpenCV DNN 的多人姿态估计 - AIUAI

OpenPose 可以对图片中单个人体目标的姿态估计,也可以处理图片中多人的姿态估计.

OpenPose 基于OpenCV DNN 的单人姿态估计 - AIUAI

这里主要记录基于 OpenCV 4.x DNN 模块和 OpenPose 开源模型的多人人体姿态估计 的实现.

采用的模型的基于 COCO 数据集训练的人体关键点模型.

1. OpenPose 网络结构

OpenPose 的多人人体姿态估计的模型结构如图:

OpenPose 基于OpenCV DNN 的多人姿态估计_第1张图片

网络输入为 hxwx3 的图片,并输出包含关键点(Keypoints)的置信图(confidence maps) 和每个关键点对(keypoint pair) 的 Part Affinity Heatmaps 的两个输出数组.

[1] - Stage 0:

采用 VGGNet 的前 10 层,提取输入图片的特征图(feature maps).

[2] - Stage 1:

采用 2-分支的 multi-stage CNN 网络结构:

(1) - 分支一

网络分支一预测人体关键点位置的 2D Confidence Maps ,如 elbow, knee, etc. 每个 Confidence Map 是一个灰度图(grayscale image),其最大值的位置坐标即为对应人体某个关键点的概率最高. 如图:

OpenPose 基于OpenCV DNN 的多人姿态估计_第2张图片

(2) - 分支二:

网络分支二预测 Part Affinities (PAF) 的2D 向量场(L, vector fields),其表示了两个关键点之间关联度(degree of association). 例如,关键点 Neck 和 Left Shoulder 之间的 Part Affinity, 如下图,属于同一个人体的关键点之间的 Affinity 值比较大.

OpenPose 基于OpenCV DNN 的多人姿态估计_第3张图片

总体来说,Confidence Maps 用于检测关键点位置;而 Affinity Maps 用于检测关键点之间的有效连接.

2. OpenPose 的 OpenCV DNN 实现

根据检测过程,主要涉及的函数有:getKeyponts()getValidPairs()getPersonwiseKeypoints() .

2.1. 模型加载

#!/usr/bin/python3
#!--*-- coding: utf-8 --*--
import cv2
import time
import numpy as np
import matplotlib.pyplot as plt


num_points = 18

keypointsMapping = ['Nose', 'Neck',
                    'R-Sho', 'R-Elb', 'R-Wr',
                    'L-Sho', 'L-Elb', 'L-Wr',
                    'R-Hip', 'R-Knee', 'R-Ank',
                    'L-Hip', 'L-Knee', 'L-Ank',
                    'R-Eye', 'L-Eye', 'R-Ear', 'L-Ear']

POSE_PAIRS = [[1,2], [1,5], [2,3], [3,4], [5,6], [6,7],
              [1,8], [8,9], [9,10], [1,11], [11,12], [12,13],
              [1,0], [0,14], [14,16], [0,15], [15,17],
              [2,17], [5,16] ]

# index of pafs correspoding to the POSE_PAIRS
# e.g for POSE_PAIR(1,2), the PAFs are located at indices (31,32) of output, Similarly, (1,5) -> (39,40) and so on.
mapIdx = [[31,32], [39,40], [33,34], [35,36], [41,42], [43,44],
          [19,20], [21,22], [23,24], [25,26], [27,28], [29,30],
          [47,48], [49,50], [53,54], [51,52], [55,56],
          [37,38], [45,46]]

colors = [[0,100,255], [0,100,255],   [0,255,255], 
          [0,100,255], [0,255,255],   [0,100,255],
          [0,255,0],   [255,200,100], [255,0,255], 
          [0,255,0],   [255,200,100], [255,0,255],
          [0,0,255],   [255,0,0],     [200,200,0], 
          [255,0,0],   [200,200,0],   [0,0,0]]

# dnn 加载模型
start = time.time()
prototxt = "./models/pose/coco/pose_deploy_linevec.prototxt"
caffemodel = "./models/pose/coco/pose_iter_440000.caffemodel"
net = cv2.dnn.readNetFromCaffe(prototxt, caffemodel)
print("[INFO]Time Taken in Model Loading: {}".format(time.time() - start))

# test img
img_cv2 = cv2.imread("test_image.jpg")
img_width, img_height = img_cv2.shape[1], img_cv2.shape[0]

# 根据长宽比,固定网路输入的 height,计算网络输入的 width.
net_height = 368
net_width = int((net_height/img_height)*img_width)

start = time.time()
in_blob = cv2.dnn.blobFromImage(
    img_cv2, 
    1.0 / 255, 
    (net_width, net_height),
    (0, 0, 0), 
    swapRB=False, 
    crop=False)

net.setInput(in_blob)
output = net.forward()
print("[INFO]Time Taken in Forward pass: {}".format(time.time() - start))

2.2. 关键点检测

2.2.1. getKeyponts( )函数

**getKeyponts( )函数功能:**对 Confidence Map 采用 NMS(Non Maximum Suppression) 来检测关键点.

def getKeypoints(probMap, threshold=0.1):
    mapSmooth = cv2.GaussianBlur(probMap,(3,3),0,0)
    mapMask = np.uint8(mapSmooth>threshold)
    
    keypoints = []
    # 1. 找出对应于关键点的所有区域的轮廓(contours)
    contours, hierarchy = cv2.findContours(
        mapMask, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
    
    # for each blob find the maxima
    # 对于每个关键点轮廓区域,找到最大值.
    for cnt in contours:
        # 2. 创建关键点的 mask;
        blobMask = np.zeros(mapMask.shape)
        blobMask = cv2.fillConvexPoly(blobMask, cnt, 1)
        # 3. 提取关键点区域的 probMap
        maskedProbMap = mapSmooth * blobMask
        # 4. 提取关键点区域的局部最大值.
        _, maxVal, _, maxLoc = cv2.minMaxLoc(maskedProbMap)
        keypoints.append(maxLoc + (probMap[maxLoc[1], maxLoc[0]],))

    return keypoints

2.2.2. 关键点 mapMask

threshold = 0.1
part = 1
probMap = output[0, part, :,:]
mapSmooth = cv2.GaussianBlur(probMap, (3, 3), 0, 0)
mapMask = np.uint8(mapSmooth > threshold)

plt.subplot(1, 3, 1)
plt.imshow(probMap)
plt.title("probMap")
plt.axis("off")
plt.subplot(1, 3, 2)
plt.imshow(mapSmooth)
plt.title("mapSmooth")
plt.axis("off")
plt.subplot(1, 3, 3)
plt.imshow(mapMask)
plt.title("mapMask")
plt.axis("off")
plt.show()

OpenPose 基于OpenCV DNN 的多人姿态估计_第4张图片

2.2.3. 关键点坐标值

threshold = 0.1
part = 1
probMap = output[0, part, :,:]
probMap = cv2.resize(probMap, (img_cv2.shape[1], img_cv2.shape[0]))
mapSmooth = cv2.GaussianBlur(probMap, (3, 3), 0, 0)
mapMask = np.uint8(mapSmooth > threshold)

keypoints = []
contours, hierarchy = cv2.findContours(mapMask, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)

for cnt in contours:
    blobMask = np.zeros(mapMask.shape)
    blobMask = cv2.fillConvexPoly(blobMask, cnt, 1)
    maskedProbMap = mapSmooth * blobMask
    _, maxVal, _, maxLoc = cv2.minMaxLoc(maskedProbMap)
    keypoints.append(maxLoc + (probMap[maxLoc[1], maxLoc[0]],))

img_cv2_copy = img_cv2.copy()
for keypoint in keypoints:
    cv2.circle(img_cv2_copy, (keypoint[0], keypoint[1]), 5, (255, 0, 0), -1, cv2.LINE_AA)

plt.figure()
plt.imshow(img_cv2_copy[:, :, ::-1])
plt.title("Keypoints")
plt.axis("off")
plt.show()

OpenPose 基于OpenCV DNN 的多人姿态估计_第5张图片

2.3. 有效关键点对检测

2.3.1. getValidPairs()函数

**getValidPairs()函数功能:**检测所有人体之间不同关键点之间的有效连接.

计算公式为:
E = ∫ u = 0 u = 1 L c ( p ( u ) ) ⋅ d j 2 − d j 1 ∣ ∣ d j 2 − d j 1 ∣ ∣ 2 d u E = \int_{u=0}^{u=1} \mathbf{L}_c (\mathbf{p}(u)) \cdot \frac{\mathbf{d}_{j2} - \mathbf{d}_{j1}}{||\mathbf{d}_{j2} - \mathbf{d}_{j1}||_2} du E=u=0u=1Lc(p(u))dj2dj12dj2dj1du
其中, p ( u ) \mathbf{p}(u) p(u) 是对两个关键点 d j 1 \mathbf{d}_{j1} dj1 d j 2 \mathbf{d}_{j2} dj2 之间的位置进行插值:
p ( u ) = ( 1 − u ) d j 1 + u d j 2 \mathbf{p}(u) = (1 - u)\mathbf{d}_{j1} + u\mathbf{d}_{j2} p(u)=(1u)dj1+udj2
L \mathbf{L} L 是Part Affinity(PAF);

d \mathbf{d} d 是两个关键点的向量.

p \mathbf{p} p 是两个关键点之间的插值点.

E \mathbf{E} E 是通过计算 PAF L \mathbf{L} L 和向量 d i j d_{ij} dij 之间的点积(dot product) 得到的.

# Find valid connections between the different 
#	joints of a all persons present
def getValidPairs(output):
    valid_pairs = []
    invalid_pairs = []
    n_interp_samples = 10
    paf_score_th = 0.1
    conf_th = 0.7
    # loop for every POSE_PAIR
    for k in range(len(mapIdx)):
        # A->B constitute a limb
        pafA = output[0, mapIdx[k][0], :, :]
        pafB = output[0, mapIdx[k][1], :, :]
        pafA = cv2.resize(pafA, (imgWidth, imgHeight))
        pafB = cv2.resize(pafB, (imgWidth, imgHeight))
		
        # 检测第一个 limb 和第二个 limb 的关键点位置
        candA = detected_keypoints[POSE_PAIRS[k][0]]
        candB = detected_keypoints[POSE_PAIRS[k][1]]
        nA = len(candA)
        nB = len(candB)
        
        # 如果检测到 joint-pair 的关键点位置,则,
        # 	检查 candA 和 candB 中每个 joint.
        # 计算两个 joints 之间的距离向量(distance vector).
        # 计算两个 joints 之间插值点集合的 PAF 值.
        # 根据上述公式,计算 score 值,判断连接的有效性.
        if( nA != 0 and nB != 0):
            valid_pair = np.zeros((0,3))
            for i in range(nA):
                max_j=-1
                maxScore = -1
                found = 0
                for j in range(nB):
                    # Find d_ij
                    d_ij = np.subtract(candB[j][:2], candA[i][:2])
                    norm = np.linalg.norm(d_ij)
                    if norm:
                        d_ij = d_ij / norm
                    else:
                        continue
                    # Find p(u)
                    interp_coord = list(
                        zip(np.linspace(candA[i][0], 
                                        candB[j][0], 
                                        num=n_interp_samples),
                            np.linspace(candA[i][1], 
                                        candB[j][1], 
                                        num=n_interp_samples)))
                    # Find L(p(u))
                    paf_interp = []
                    for k in range(len(interp_coord)):
                        paf_interp.append(
                            [pafA[int(round(interp_coord[k][1])), 
                                  int(round(interp_coord[k][0]))],
                             pafB[int(round(interp_coord[k][1])), 
                                  int(round(interp_coord[k][0]))] 
                            ]) 
                    # Find E
                    paf_scores = np.dot(paf_interp, d_ij)
                    avg_paf_score = sum(paf_scores)/len(paf_scores)
                    
                    # 判断连接是否有效.
                    # 如果对应于 PAF 的插值向量值大于阈值,则连接有效.
                    if (len(np.where(paf_scores > paf_score_th)[0])/
                        n_interp_samples ) > conf_th :
                        if avg_paf_score > maxScore:
                            max_j = j
                            maxScore = avg_paf_score
                            found = 1
                # Append the connection to the list
                if found:            
                    valid_pair = np.append(
                        valid_pair, 
                        [[candA[i][3], candB[max_j][3], maxScore]], 
                        axis=0)

            # Append the detected connections to the global list
            valid_pairs.append(valid_pair)
        else: # If no keypoints are detected
            print("No Connection : k = {}".format(k))
            invalid_pairs.append(k)
            valid_pairs.append([])
    print(valid_pairs)
    
    return valid_pairs, invalid_pairs

2.3.2. 关键点对检测具体实现分析

有效的关键点对(joint pair) 是指两个关键点的连接,属于相同的人体.

最简单的方式是,计算一个关键点与其它所有可能的关键点之间的最小距离,来判断关键点对的有效性.

例如下图,可以计算 Nose 关键点与其它所有 Necks 关键点之间的距离,最小距离的 Neck 关键点,则对应于同一个人体.

OpenPose 基于OpenCV DNN 的多人姿态估计_第6张图片

但是,这种方法不是对所有的关键点对都是有效的,尤其是对与图像中包含很多人体,或者有关键点缺失的时候.

例如下图,对于关键点对, Left-Elbow -> Left Wrist,第三个人体的 wrist 关键点与第二个人体的 elbow 关键点的距离,比与其同一人体的 elbow 关键点的距离更小. 但该关键点对并不是有效的.

OpenPose 基于OpenCV DNN 的多人姿态估计_第7张图片

而 Part Affinity Maps 的作用是,给定沿着两个关键点对的仿射(affinity)的方向. 因此,有效的关键点对不仅具有最小的距离,其方向也应该顺着 PAF Heatmaps 方向.

例如,Left-Elbow -> Left Wrist 连接的 Heatmap,如下图,即使最小距离的关键点是错误的,但,由于 PAF 只能顺着 Elbow 和 Wrist 的单位向量,所以能正确检测有效的关键点对:

OpenPose 基于OpenCV DNN 的多人姿态估计_第8张图片

OpenPose 中采用的方法为:

[1] - 将关键点对的两个点之间的连线进行划分,得到该连线上的 n 个点(Divide the line joining the two points comprising the pair. Find n points on this line.);

[2] - 判断这些点上的 PAF 是否与连接该关键点的线的方向相同(Check if the PAF on these points have the same direction as that of the line joining the points for this pair);

[3] - 如果方向满足特定程度,则为有效的关键点对(If the direction matches to a certain extent, then it is valid pair.)

代码实现与分析:

对于每个关键点对(body part pair, keypoints pair),

[1] - 选择属于同一个关键点对的关键点. 并分别存放在两个列表: candAcandB. candA 列表中的每个关键点可以与 candB 中的某些关键点相连接. 如下图,给出了 candAcandB 中的 Neck -> Right-Shoulder 关键点对的所有关键点:

OpenPose 基于OpenCV DNN 的多人姿态估计_第9张图片

对应 Python 实现:

pafA = output[0, mapIdx[k][0], :, :]
pafB = output[0, mapIdx[k][1], :, :]
pafA = cv2.resize(pafA, (frameWidth, frameHeight))
pafB = cv2.resize(pafB, (frameWidth, frameHeight))
 
# Find the keypoints for the first and second limb
candA = detected_keypoints[POSE_PAIRS[k][0]]
candB = detected_keypoints[POSE_PAIRS[k][1]]

[2] - 计算两个关键点之间的单位向量,其给定了关节点之间连线的方向.

d_ij = np.subtract(candB[j][:2], candA[i][:2])
norm = np.linalg.norm(d_ij)
if norm:
    d_ij = d_ij / norm

[3] - 计算两个关键点之间连线的 10 个插值点.

# Find p(u)
interp_coord = list(zip(np.linspace(candA[i][0], candB[j][0], num=n_interp_samples),
                        np.linspace(candA[i][1], candB[j][1], num=n_interp_samples)))
# Find L(p(u))
paf_interp = []
for k in range(len(interp_coord)):
    paf_interp.append(
        [pafA[int(round(interp_coord[k][1])), 
              int(round(interp_coord[k][0]))],
         pafB[int(round(interp_coord[k][1])), 
              int(round(interp_coord[k][0]))] 
        ]) 

[4] - 计算插值点的 PAF 与单位向量 d_ij 之间的点积(dot product).

# Find E
paf_scores = np.dot(paf_interp, d_ij)
avg_paf_score = sum(paf_scores)/len(paf_scores)

[5] - 如果这些插值点的 70% 的都满足判定标准,则该关键点对是有效的.

# Check if the connection is valid
# If the fraction of interpolated vectors 
#	aligned with PAF is higher then threshold -> Valid Pair  
if (len(np.where(paf_scores > paf_score_th)[0])/n_interp_samples ) > conf_th :
    if avg_paf_score > maxScore:
        max_j = j
        maxScore = avg_paf_score

2.4. 同一人体关键点的组合

2.4.1. getPersonwiseKeypoints() 函数

**getPersonwiseKeypoints()函数功能:**计算得到属于每个人体的关键点集合.

对于每个检测到的有效 joint pair,分配属于一个人体的 joints.

def getPersonwiseKeypoints(valid_pairs, invalid_pairs):
    # 每一行的最后一个值为 overall score.
    personwiseKeypoints = -1 * np.ones((0, 19))

    for k in range(len(mapIdx)):
        if k not in invalid_pairs:
            partAs = valid_pairs[k][:,0]
            partBs = valid_pairs[k][:,1]
            indexA, indexB = np.array(POSE_PAIRS[k])

            for i in range(len(valid_pairs[k])): 
                found = 0
                person_idx = -1
                for j in range(len(personwiseKeypoints)):
                    if personwiseKeypoints[j][indexA] == partAs[i]:
                        person_idx = j
                        found = 1
                        break

                if found:
                    personwiseKeypoints[person_idx][indexB] = partBs[i]
                    personwiseKeypoints[person_idx][-1] += 
                    	keypoints_list[partBs[i].astype(int), 2] + 
                        valid_pairs[k][i][2]

                # if find no partA in the subset, create a new subset
                elif not found and k < 17:
                    row = -1 * np.ones(19)
                    row[indexA] = partAs[i]
                    row[indexB] = partBs[i]
                    # add the keypoint_scores for the two keypoints and the paf_score 
                    row[-1] = sum(
                        keypoints_list[valid_pairs[k][i,:2].astype(int), 2]) 
                    + valid_pairs[k][i][2]
                    personwiseKeypoints = np.vstack([personwiseKeypoints, row])
    return personwiseKeypoints

2.4.2. 关键点组合的具体实现分析

计算得到所有关键点之间的关键点对后,可以将具有相同关键点检测候选值的关键点对,组合为多人的姿态估计.

代码实现具体分析

[1] - 首先创建保存每个人体的所有关键点的空列表.

然后对每个关键点对,判断 partA 是否已经在列表里, 如果已经在列表里,则表示该关键点对属于该列表,且 partB 也属于同一人体. 因此,添加该关键点对的 partB 到 partA 所在的列表.

# 空列表
personwiseKeypoints = -1 * np.ones((0, 19))
for j in range(len(personwiseKeypoints)):
    if personwiseKeypoints[j][indexA] == partAs[i]:
        person_idx = j
        found = 1
        break
 
if found:
    personwiseKeypoints[person_idx][indexB] = partBs[i]

[2] - 如果 partA 不在任一人体列表里,则表示该关键点对属于一个新出现的人体,故创建新的列表.

# if find no partA in the subset, create a new subset
elif not found and k < 17:
    row = -1 * np.ones(19)
    row[indexA] = partAs[i]
    row[indexB] = partBs[i]

2.5. 姿态检测结果可视化

##
img_cv2_copy = img_cv2.copy()
for i in range(num_points):
    for j in range(len(detected_keypoints[i])):
        cv2.circle(img_cv2_copy, 
                   detected_keypoints[i][j][0:2], 
                   5, 
                   colors[i], 
                   -1, 
                   cv2.LINE_AA)

for i in range(17):
    for n in range(len(personwiseKeypoints)):
        index = personwiseKeypoints[n][np.array(POSE_PAIRS[i])]
        if -1 in index:
            continue
        B = np.int32(keypoints_list[index.astype(int), 0])
        A = np.int32(keypoints_list[index.astype(int), 1])
        cv2.line(img_cv2_copy, 
                 (B[0], A[0]), (B[1], A[1]), 
                 colors[i], 
                 3, 
                 cv2.LINE_AA)
        
plt.figure()
plt.imshow(img_cv2_copy[:, :, ::-1])
plt.title("Results")
plt.axis("off")
plt.show()

如:

OpenPose 基于OpenCV DNN 的多人姿态估计_第10张图片

3. 完整实现

模型下载:

COCO: http://posefs1.perception.cs.cmu.edu/OpenPose/models/pose/coco/pose_iter_440000.caffemodel

#!/usr/bin/python3
#!--*-- coding: utf-8 --*--
import os
import cv2
import time
import numpy as np
import matplotlib.pyplot as plt


class general_mulitpose_model(object):
    def __init__(self):
        self.point_names = ['Nose', 'Neck',
                            'R-Sho', 'R-Elb', 'R-Wr',
                            'L-Sho', 'L-Elb', 'L-Wr',
                            'R-Hip', 'R-Knee', 'R-Ank',
                            'L-Hip', 'L-Knee', 'L-Ank',
                            'R-Eye', 'L-Eye', 'R-Ear', 'L-Ear']
        self.point_pairs = [[1,2], [1,5], [2,3], [3,4], [5,6], [6,7],
                            [1,8], [8,9], [9,10], [1,11], [11,12], [12,13],
                            [1,0], [0,14], [14,16], [0,15], [15,17],
                            [2,17], [5,16] ]

        # index of pafs correspoding to the self.point_pairs
        # e.g for point_pairs(1,2), the PAFs are located at indices (31,32) of output,
        #   Similarly, (1,5) -> (39,40) and so on.
        self.map_idx = [[31,32], [39,40], [33,34], [35,36], [41,42], [43,44],
                        [19,20], [21,22], [23,24], [25,26], [27,28], [29,30],
                        [47,48], [49,50], [53,54], [51,52], [55,56],
                        [37,38], [45,46]]

        self.colors = [[0,100,255], [0,100,255],   [0,255,255],
                       [0,100,255], [0,255,255],   [0,100,255],
                       [0,255,0],   [255,200,100], [255,0,255],
                       [0,255,0],   [255,200,100], [255,0,255],
                       [0,0,255],   [255,0,0],     [200,200,0],
                       [255,0,0],   [200,200,0],   [0,0,0]]

        self.num_points = 18
        self.pose_net = self.get_model()


    def get_model(self):
        prototxt = "./models/pose/coco/pose_deploy_linevec.prototxt"
        caffemodel = "./models/pose/coco/pose_iter_440000.caffemodel"
        coco_net = cv2.dnn.readNetFromCaffe(prototxt, caffemodel)

        return coco_net

    def getKeypoints(self, probMap, threshold=0.1):
        mapSmooth = cv2.GaussianBlur(probMap, (3, 3), 0, 0)
        mapMask = np.uint8(mapSmooth > threshold)

        keypoints = []
        # find the blobs
        contours, hierarchy = cv2.findContours(mapMask,
                                          cv2.RETR_TREE,
                                          cv2.CHAIN_APPROX_SIMPLE)

        # for each blob find the maxima
        for cnt in contours:
            blobMask = np.zeros(mapMask.shape)
            blobMask = cv2.fillConvexPoly(blobMask, cnt, 1)
            maskedProbMap = mapSmooth * blobMask
            _, maxVal, _, maxLoc = cv2.minMaxLoc(maskedProbMap)
            keypoints.append(maxLoc + (probMap[maxLoc[1], maxLoc[0]],))

        return keypoints

    def getValidPairs(self, output, detected_keypoints, img_width, img_height):
        valid_pairs = []
        invalid_pairs = []
        n_interp_samples = 10
        paf_score_th = 0.1
        conf_th = 0.7

        for k in range(len(self.map_idx)):
            # A->B constitute a limb
            pafA = output[0, self.map_idx[k][0], :, :]
            pafB = output[0, self.map_idx[k][1], :, :]
            pafA = cv2.resize(pafA, (img_width, img_height))
            pafB = cv2.resize(pafB, (img_width, img_height))

            # Find the keypoints for the first and second limb
            candA = detected_keypoints[self.point_pairs[k][0]]
            candB = detected_keypoints[self.point_pairs[k][1]]
            nA = len(candA)
            nB = len(candB)

            if (nA != 0 and nB != 0):
                valid_pair = np.zeros((0, 3))
                for i in range(nA):
                    max_j = -1
                    maxScore = -1
                    found = 0
                    for j in range(nB):
                        # Find d_ij
                        d_ij = np.subtract(candB[j][:2], candA[i][:2])
                        norm = np.linalg.norm(d_ij)
                        if norm:
                            d_ij = d_ij / norm
                        else:
                            continue
                        # Find p(u)
                        interp_coord = list(
                            zip(np.linspace(candA[i][0], candB[j][0], num=n_interp_samples),
                                np.linspace(candA[i][1], candB[j][1], num=n_interp_samples)))
                        # Find L(p(u))
                        paf_interp = []
                        for k in range(len(interp_coord)):
                            paf_interp.append([pafA[int(round(interp_coord[k][1])), int(
                                round(interp_coord[k][0]))],
                                               pafB[int(round(interp_coord[k][1])), int(
                                                   round(interp_coord[k][0]))]])
                        # Find E
                        paf_scores = np.dot(paf_interp, d_ij)
                        avg_paf_score = sum(paf_scores) / len(paf_scores)

                        # Check if the connection is valid
                        if (len(np.where(paf_scores > paf_score_th)[
                                    0]) / n_interp_samples) > conf_th:
                            if avg_paf_score > maxScore:
                                max_j = j
                                maxScore = avg_paf_score
                                found = 1

                    # Append the connection to the list
                    if found:
                        valid_pair = np.append(valid_pair,
                                               [[candA[i][3], candB[max_j][3], maxScore]], axis=0)

                # Append the detected connections to the global list
                valid_pairs.append(valid_pair)
            else:  # If no keypoints are detected
                print("No Connection : k = {}".format(k))
                invalid_pairs.append(k)
                valid_pairs.append([])

        return valid_pairs, invalid_pairs


    def getPersonwiseKeypoints(self, valid_pairs, invalid_pairs, keypoints_list):
        personwiseKeypoints = -1 * np.ones((0, 19))

        for k in range(len(self.map_idx)):
            if k not in invalid_pairs:
                partAs = valid_pairs[k][:, 0]
                partBs = valid_pairs[k][:, 1]
                indexA, indexB = np.array(self.point_pairs[k])

                for i in range(len(valid_pairs[k])):
                    found = 0
                    person_idx = -1
                    for j in range(len(personwiseKeypoints)):
                        if personwiseKeypoints[j][indexA] == partAs[i]:
                            person_idx = j
                            found = 1
                            break

                    if found:
                        personwiseKeypoints[person_idx][indexB] = partBs[i]
                        personwiseKeypoints[person_idx][-1] += keypoints_list[
                                                                   partBs[i].astype(int), 2] + \
                                                               valid_pairs[k][i][2]

                    # if find no partA in the subset, create a new subset
                    elif not found and k < 17:
                        row = -1 * np.ones(19)
                        row[indexA] = partAs[i]
                        row[indexB] = partBs[i]
                        # add the keypoint_scores for the two keypoints and the paf_score
                        row[-1] = sum(keypoints_list[valid_pairs[k][i, :2].astype(int), 2]) + \
                                  valid_pairs[k][i][2]
                        personwiseKeypoints = np.vstack([personwiseKeypoints, row])

        return personwiseKeypoints


    def predict(self, imgfile):
        img_cv2 = cv2.imread(imgfile)
        img_width, img_height = img_cv2.shape[1], img_cv2.shape[0]

        net_height = 368
        net_width = int((net_height / img_height) * img_width)

        start = time.time()
        in_blob = cv2.dnn.blobFromImage(
            img_cv2,
            1.0 / 255,
            (net_width, net_height),
            (0, 0, 0),
            swapRB=False,
            crop=False)

        self.pose_net.setInput(in_blob)
        output = self.pose_net.forward()
        print("[INFO]Time Taken in Forward pass: {}".format(time.time() - start))

        detected_keypoints = []
        keypoints_list = np.zeros((0, 3))
        keypoint_id = 0
        threshold = 0.1
        for part in range(self.num_points):
            probMap = output[0, part, :, :]
            probMap = cv2.resize(probMap, (img_cv2.shape[1], img_cv2.shape[0]))
            keypoints = self.getKeypoints(probMap, threshold)
            print("Keypoints - {} : {}".format(self.point_names[part], keypoints))
            keypoints_with_id = []
            for i in range(len(keypoints)):
                keypoints_with_id.append(keypoints[i] + (keypoint_id,))
                keypoints_list = np.vstack([keypoints_list, keypoints[i]])
                keypoint_id += 1

            detected_keypoints.append(keypoints_with_id)

        valid_pairs, invalid_pairs = \
        	self.getValidPairs(output, 
                               detected_keypoints, 
                               img_width, 
                               img_height)
        personwiseKeypoints = \
        	self.getPersonwiseKeypoints(valid_pairs, 
                                        invalid_pairs, 
                                        keypoints_list)

        return personwiseKeypoints, keypoints_list


    def vis_pose(self, img_file, personwiseKeypoints, keypoints_list):
        img_cv2 = cv2.imread(img_file)
        for i in range(17):
            for n in range(len(personwiseKeypoints)):
                index = personwiseKeypoints[n][np.array(self.point_pairs[i])]
                if -1 in index:
                    continue
                B = np.int32(keypoints_list[index.astype(int), 0])
                A = np.int32(keypoints_list[index.astype(int), 1])
                cv2.line(img_cv2, (B[0], A[0]), (B[1], A[1]), self.colors[i], 3, cv2.LINE_AA)

        plt.figure()
        plt.imshow(img_cv2[:, :, ::-1])
        plt.title("Results")
        plt.axis("off")
        plt.show()


if __name__ == '__main__':
    print("[INFO]MultiPose estimation.")
    img_file = "multipose_test_image.jpg"
    
    start = time.time()
    multipose_model = general_mulitpose_model()
    print("[INFO]Time Taken in Model Loading: {}".\
          format(time.time() - start))
    personwiseKeypoints, keypoints_list = \
    		multipose_model.predict(img_file)
    multipose_model.vis_pose(img_file, 
                             personwiseKeypoints, 
                             keypoints_list)
    print(personwiseKeypoints)
    print(keypoints_list)
    print("[INFO]Done.")

如:

OpenPose 基于OpenCV DNN 的多人姿态估计_第11张图片

4. 参考资料

[1] - Multi-Person Pose Estimation in OpenCV using OpenPose - 2018.09.11

[2] - Github - spmallick/learnopencv/OpenPose-Multi-Person

你可能感兴趣的:(姿态估计)