def body_seg(filename="./pic/test.jpg", savefilename="./pic/fore.jpg"):
APP_ID = '19037846'
API_KEY = 'pi8i47A8jGBH1VZDr8ioqgfC'
SECRET_KEY = 'qwcqweVfjf9pX2fB6MykU15u6XszXko4'
client = AipBodyAnalysis(APP_ID, API_KEY, SECRET_KEY)
image = get_file(filename)
options = {}
options["type"] = "foreground"
res = client.bodySeg(image, options)
foreground = base64.b64decode(res["foreground"])
with open(savefilename, "wb") as f:
f.write(foreground)
def AutoCutout(path):
APP_ID = '19815134'
API_KEY = 'i7yBb3Fx3e4ZMILo8nHsrZQT'
SECRET_KEY = 'wuEXaPYbz1RXlAdDYYRj49tlPoNZdqfW'
client = AipBodyAnalysis(APP_ID, API_KEY, SECRET_KEY)
imgfile = path
ori_img = cv2.imread(imgfile)
height, width, _ = ori_img.shape
with open(imgfile, 'rb') as fp:
img_info = fp.read()
seg_res = client.bodySeg(img_info)
labelmap = base64.b64decode(seg_res['labelmap'])
nparr = np.fromstring(labelmap, np.uint8)
labelimg = cv2.imdecode(nparr, 1)
labelimg = cv2.resize(labelimg, (width, height), interpolation=cv2.INTER_NEAREST)
new_img = np.where(labelimg == 1, 255, labelimg)
result = cv2.bitwise_and(ori_img, new_img)
cv2.imwrite(path, result)
def Seg_img(jpg_path, crop_path, save_path):
'''
调用API 分割人像图片
:param jpg_path: 原图像
:param crop_path: 裁剪之后图像
:param save_path: 生成 mask 存放路径
:return:
'''
# 通过百度控制台申请得到的 AK 和 SK；
APP_ID = "23633750"
API_KEY = 'uqnHjMZfChbDHvPqWgjeZHCR'
SECRET_KEY = 'KIKTgD5Dh0EGyqn74Zqq8wHWB39uKaS3'
client = AipBodyAnalysis(APP_ID, API_KEY, SECRET_KEY)
# 文件夹
jpg_file = os.listdir(jpg_path)
# 要保存的文件夹
for i in jpg_file:
open_file = os.path.join(jpg_path, i)
save_file = os.path.join(save_path, i)
if not os.path.exists(save_file): #文件不存在时，进行下步操作
img = cv2.imread(open_file) # 获取图像尺寸
height, width, _ = img.shape
if crop_path: # 若Crop_path 不为 None,则不进行裁剪
crop_file = os.path.join(crop_path, i)
img = img[100:-1, 300:-400] #图片太大，对图像进行裁剪
cv2.imwrite(crop_file, img)
image = get_file_content(crop_file)
else:
image = get_file_content(open_file)
res = client.bodySeg(image) #调用百度API 对人像进行分割
labelmap = base64.b64decode(res['labelmap'])
labelimg = np.frombuffer(labelmap, np.uint8) # 转化为np数组 0-255
labelimg = cv2.imdecode(labelimg, 1)
labelimg = cv2.resize(labelimg, (width, height),
interpolation=cv2.INTER_NEAREST)
img_new = np.where(labelimg == 1, 255, labelimg) # 将 1 转化为 255
cv2.imwrite(save_file, img_new)
print(save_file, 'save successfully')
def cut_picture(src):
# 在百度云中申请，每天各接口有 500 次调用限制.
APP_ID = '16628525'
API_KEY = '5ioBzjijPln33f7mPzyProbc'
SECRET_KEY = 'FR4URtsr2Q77r6RvNyld4swls1Eik5Pu'
client = AipBodyAnalysis(APP_ID, API_KEY, SECRET_KEY)
# 读取待分割人像图片
imgfile = src
ori_img = cv2.imread(imgfile)
height, width, _ = ori_img.shape
with open(imgfile, 'rb') as fp:
img_info = fp.read()
# print("img_info:", img_info)
seg_res = client.bodySeg(img_info)
labelmap = base64.b64decode(seg_res['labelmap'])
nparr = np.fromstring(labelmap, np.uint8)
labelimg = cv2.imdecode(nparr, 1)
labelimg = cv2.resize(labelimg, (width, height),
interpolation=cv2.INTER_NEAREST)
new_img = np.where(labelimg == 1, 255, labelimg)
# maskfile = imgfile.replace('.jpg', '_mask.png')
maskfile_dir = './static/image/cut/' + \
imgfile.split('/')[-1].split('.')[0]+'_mask.png'
cv2.imwrite(maskfile_dir, new_img)
# res_imgfile = imgfile.replace('.jpg', '_res.jpg')
result = cv2.bitwise_and(ori_img, new_img)
# 保存最终分割出的人像图片到原始图片所在目录
result_path = './static/image/cut/' + \
imgfile.split('/')[-1].split('.')[0]+'_res.jpg'
cv2.imwrite(result_path, result)
def bodyseg(filename):
# APP_ID = 'd90755ad2f2047dbabb12ad0adaa0b03'
# API_KEY = 'b7b0f4d01f7f4aef9b5453f6558c23b1'
# SECRET_KEY = '6ad666162ef24213b5bde7bdd904fcbe'
client = AipBodyAnalysis(APP_ID, API_KEY, SECRET_KEY)
# """ 读取图片 """
image = get_file_content(filename)
res = client.bodySeg(image)
print(res)
labelmap = base64.b64decode(res['labelmap'])
# time.sleep(2)
nparr_labelmap = np.fromstring(labelmap, np.uint8)
labelmapimg = cv2.imdecode(nparr_labelmap, 1)
print(labelmapimg.shape)
im_new_labelmapimg = np.where(labelmapimg == 1, 255, labelmapimg)
# print(im_new_labelmapimg.shape)
# img=cv2.cvtColor(im_new_labelmapimg, cv2.COLOR_BGR2GRAY)
# return img.astype('uint8')
# cv2.imwrite('outline.png',im_new_labelmapimg)
return im_new_labelmapimg
# 在百度云中申请，每天各接口有 500 次调用限制.
startTime = time.time()
APP_ID = '19037846'
API_KEY = 'pi8i47A8jGBH1VZDr8ioqgfC'
SECRET_KEY = 'qwcqweVfjf9pX2fB6MykU15u6XszXko4'
client = AipBodyAnalysis(APP_ID, API_KEY, SECRET_KEY)
imgfile = 'test.jpg'
ori_img = cv2.imread(imgfile)
height, width, _ = ori_img.shape
with open(imgfile, 'rb') as fp:
img_info = fp.read()
seg_res = client.bodySeg(img_info)
labelmap = base64.b64decode(seg_res['labelmap'])
nparr = np.fromstring(labelmap, np.uint8)
labelimg = cv2.imdecode(nparr, 1)
labelimg = cv2.resize(labelimg, (width, height),
interpolation=cv2.INTER_NEAREST)
new_img = np.where(labelimg == 1, 255, labelimg)
maskfile = imgfile.replace('.jpg', '_mask.png')
cv2.imwrite(maskfile, new_img)
res_imgfile = imgfile.replace('.jpg', '_res.jpg')
result = cv2.bitwise_and(ori_img, new_img)
cv2.imwrite(res_imgfile, result)
endTime = time.time()
print("useTime: ", endTime - startTime)
print('Done.')
APP_ID = '22952720'
API_KEY = 'inMk4SrpYUxPmSUGmqGw89Ow'
SECRET_KEY = '7ecoXWQBLwwEjwp83Od0jV5EZM166oG7'
client = AipBodyAnalysis(APP_ID, API_KEY, SECRET_KEY)
""" 读取图片 """
def get_file_content(filePath):
with open(filePath, 'rb') as fp:
return fp.read()
image = get_file_content('post_img.jpg')
""" 调用人像分割 """
client.bodySeg(image)
""" 如果有可选参数 """
options = {}
options["type"] = "labelmap"
def base64_to_rgb(base64_str):
if isinstance(base64_str, bytes):
base64_str = base64_str.decode("utf-8")
img_data = base64.b64decode(base64_str)
return skimage.io.imread(img_data, plugin='imageio')
""" 带参数调用人像分割 """
resp = client.bodySeg(image, options)
APP_ID = '14724666'
API_KEY = '9RgncDsMemqT2vFNlPGWsVRM'
SECRET_KEY = 'RqQh1PEKh3REH0cNgBwKCBZGRzeaitsP'
client = AipBodyAnalysis(APP_ID, API_KEY, SECRET_KEY)
""" 读取图片 """
def get_file_content(filePath):
with open(filePath, 'rb') as fp:
return fp.read()
image = get_file_content('example.png')
""" 调用人像分割 """
dict = client.bodySeg(image)
labelmap = dict['labelmap']
print(labelmap)
ret = base64.b64decode(labelmap)
print(ret)
with open('labelmap.png', 'wb') as fp:
fp.write(ret)
mask = cv.imread('labelmap.png')
mask = cv.resize(mask, (640, 360), cv.INTER_NEAREST)
print(mask.shape)
print(np.max(mask))
print(np.min(mask))
mask = mask * 255
class BodyClient(object):
# APP_ID = '44eab39f0ed44844b94f487a6e88fdbc' # 'd90755ad2f2047dbabb12ad0adaa0b03'
# API_KEY = '55e735d6888b46908915f3533a6b7442' # '22f1025b88f54494bcf5d980697b4b83 '
# SECRET_KEY = '41213cbdaffa483d9ed9a59a24157d4b' # '4a4b41139c204905be1db659d751355f'
APP_ID = '18176250' # 'd90755ad2f2047dbabb12ad0adaa0b03'
API_KEY = 'jgfFvfLmKWBTuuIFdFgF4YXI' # '22f1025b88f54494bcf5d980697b4b83 '
SECRET_KEY = 'NCnPnPOcSRexlxoz9fI75fUmEGl3H15f' # '4a4b41139c204905be1db659d751355f'
def __init__(self):
self.client = AipBodyAnalysis(BodyClient.APP_ID, BodyClient.API_KEY, BodyClient.SECRET_KEY)
def _body_seg(self,filename):
image = get_file_content(filename)
res = self.client.bodySeg(image)
labelmap = base64.b64decode(res['labelmap'])
nparr_labelmap = np.fromstring(labelmap, np.uint8)
labelmapimg = cv2.imdecode(nparr_labelmap, 1)
im_new_labelmapimg = np.where(labelmapimg == 1, 255, labelmapimg)
return im_new_labelmapimg
def _canny_points(self,img,rect=None):
if rect:
x,y,w,h = rect
else:
x,y=0,0
h,w = img.shape[:2]
body = img[y:y + h, x:x + w]
edges = cv2.Canny(body, 10, 100)
edgesmat = np.mat(edges)
points = [(j + x, i + y) for i in range(h) for j in range(w) if edgesmat[i, j] == 255]
return points
def body_seg_points(self,filename):
img_seg=self._body_seg(filename)
return self._canny_points(img_seg)
def body_seg(self,filename):
img_seg = self._body_seg(filename)
h, w = img_seg.shape[:2]
return self._canny_points(img_seg),w,h
'''
{'person_num': 2, 'person_info': [{'body_parts': {'left_hip': {'y': 549.78125, 'x': 423.34375, 'score': 0.8641700744628906}, 'top_head': {'y': 295.46875, 'x': 394.0, 'score': 0.8867737650871277}, 'right_mouth_corner': {'y': 344.375, 'x': 384.21875, 'score': 0.8865712285041809}, 'neck': {'y': 363.9375, 'x': 394.0, 'score': 0.8912984728813171}, 'left_shoulder': {'y': 383.5, 'x': 442.90625, 'score': 0.8800243139266968}, 'left_knee': {'y': 657.375, 'x': 433.125, 'score': 0.8804177045822144}, 'left_ankle': {'y': 755.1875, 'x': 423.34375, 'score': 0.8549085855484009}, 'left_mouth_corner': {'y': 344.375, 'x': 403.78125, 'score': 0.8695278763771057}, 'right_elbow': {'y': 442.1875, 'x': 305.96875, 'score': 0.9053295850753784}, 'right_ear': {'y': 324.8125, 'x': 374.4375, 'score': 0.8913755416870117}, 'nose': {'y': 324.8125, 'x': 394.0, 'score': 0.8767616748809814}, 'left_eye': {'y': 315.03125, 'x': 403.78125, 'score': 0.8842508792877197}, 'right_eye': {'y': 315.03125, 'x': 384.21875, 'score': 0.872444748878479}, 'right_hip': {'y': 549.78125, 'x': 374.4375, 'score': 0.8706536293029785}, 'left_wrist': {'y': 491.09375, 'x': 462.46875, 'score': 0.8681846857070923}, 'left_ear': {'y': 324.8125, 'x': 413.5625, 'score': 0.8833358883857727}, 'left_elbow': {'y': 432.40625, 'x': 491.8125, 'score': 0.8757244944572449}, 'right_shoulder': {'y': 383.5, 'x': 345.09375, 'score': 0.8604100942611694}, 'right_ankle': {'y': 755.1875, 'x': 364.65625, 'score': 0.883700966835022}, 'right_knee': {'y': 657.375, 'x': 364.65625, 'score': 0.8726198673248291}, 'right_wrist': {'y': 491.09375, 'x': 335.3125, 'score': 0.8524751663208008}}, 'location': {'height': 522.3967895507812, 'width': 213.4878540039062, 'top': 279.5125427246094, 'score': 0.9985131025314331, 'left': 288.0614013671875}}, {'body_parts': {'left_hip': {'y': 539.0, 'x': 413.25, 'score': 0.2676204741001129}, 'top_head': {'y': 741.5, 'x': 524.625, 'score': 0.0297189150005579}, 'right_mouth_corner': {'y': 478.25, 'x': 463.875, 'score': 0.009633682668209076}, 'neck': {'y': 852.875, 'x': 332.25, 'score': 0.01634016819298267}, 'left_shoulder': {'y': 377.0, 'x': 423.375, 'score': 0.0272684283554554}, 'left_knee': {'y': 650.375, 'x': 423.375, 'score': 0.3098172545433044}, 'left_ankle': {'y': 751.625, 'x': 433.5, 'score': 0.4415453672409058}, 'left_mouth_corner': {'y': 478.25, 'x': 463.875, 'score': 0.01229123119264841}, 'right_elbow': {'y': 427.625, 'x': 494.25, 'score': 0.08809270709753036}, 'right_ear': {'y': 680.75, 'x': 750, 'score': 0.02279716171324253}, 'nose': {'y': 488.375, 'x': 453.75, 'score': 0.02511453814804554}, 'left_eye': {'y': 488.375, 'x': 443.625, 'score': 0.02269705384969711}, 'right_eye': {'y': 751.625, 'x': 750, 'score': 0.02191649936139584}, 'right_hip': {'y': 539.0, 'x': 372.75, 'score': 0.1868444383144379}, 'left_wrist': {'y': 488.375, 'x': 474.0, 'score': 0.3365231156349182}, 'left_ear': {'y': 893.375, 'x': 403.125, 'score': 0.007937739603221416}, 'left_elbow': {'y': 437.75, 'x': 484.125, 'score': 0.1944440901279449}, 'right_shoulder': {'y': 377.0, 'x': 433.5, 'score': 0.02875573188066483}, 'right_ankle': {'y': 751.625, 'x': 423.375, 'score': 0.1604309529066086}, 'right_knee': {'y': 670.625, 'x': 372.75, 'score': 0.1398747861385345}, 'right_wrist': {'y': 488.375, 'x': 474.0, 'score': 0.07319473475217819}}, 'location': {'height': 539.47509765625, 'width': 126.1507263183594, 'top': 458.58251953125, 'score': 0.00636356882750988, 'left': 622.8492431640625}}], 'log_id': 1953527121404955486}
'''
def _body_part(self,filename):
image = get_file_content(filename)
para = self.client.bodyAnalysis(image)
if DEBUG:
print(para)
person_num=para.get('person_num',0)
if person_num < 1:
raise NoBodyException("文件%s 没有检测到人像，详细信息：%s"%(filename,para))
person=para['person_info'][0]
# score = person['location']['score']
# if score < 0.5:
# raise NoBodyException()
loc = person['location']
x_left = int(loc['left'])
y_top = int(loc['top'])
w = int(loc['width'])
h = int(loc['height'])
return person['body_parts'],(x_left,y_top,w,h)
#top_head, left_ankle, right_ankle
def body_points(self,filename):
parts,rect = self._body_part(filename)
points = {k: (v['x'], v['y']) for k, v in parts.items()}
return points,rect
def process_body(self,filename):
img = cv2.imread(filename)
height, width = img.shape[:2]
body_points, rect = self.body_points(filename)
x_left, y_top, w, h = rect
new_rect=expand_rect(x_left,y_top,w,h,width,height)
img_seg = self._body_seg(filename)
# cv2.imshow("seg",img_seg)
outline_points = self._canny_points(img_seg, new_rect)
# print(outline_points)
return body_points, outline_points, rect
class BaiduApiUtil:
def __init__(self, configPath, configSection):
keyInfo = fu.readConfig(configPath, configSection)
self.appid = keyInfo['appid']
self.api_key = keyInfo['api_key']
self.secret_key = keyInfo['secret_key']
self.client = AipBodyAnalysis(self.appid, self.api_key, self.secret_key)
self.filename = None
self.picture = None
self.picture_size = None
self.picture_format = None
def upload(self, filename):
self.picture = fu.readPicture(filename)
self.filename = filename.split('.')[0]
img = Image.open(filename)
self.picture_size = img.size
self.picture_format = img.format
def getAccessToken(self, configPath, configSection):
keyInfo = fu.readConfig(configPath, configSection)
host = keyInfo['addr'] % (keyInfo['grant_type'], keyInfo['client_id'], keyInfo['client_secret'])
response = requests.post(host)
if response.status_code != 200:
print('Error Happened When Acquiring Access Token')
return -1
content = demjson.decode(response.text)
if 'error' in content.keys():
print('Invalid API Key or Secret Key')
return -1
return content['refresh_token']
def getBodyAnalysis(self):
response = self.client.bodyAnalysis(self.picture)
if 'error_code' in response.keys():
print(response['error_msg'])
exit(-1)
return response
def getBodySeg(self):
result = self.client.bodySeg(self.picture, {'type':'labelmap'})
# foreground = base64.b64decode(result['foreground'])
labelmap = base64.b64decode(result['labelmap'])
# scoremap = base64.b64decode(result['scoremap'])
# nparr_foreground = np.fromstring(foreground, np.uint8)
# foregroundimg = cv2.imdecode(nparr_foreground, 1)
# foregroundimg = cv2.resize(foregroundimg, self.picture_size, interpolation=cv2.INTER_NEAREST)
# im_new_foreground = np.where(foregroundimg == 1, 10, foregroundimg)
# cv2.imwrite(self.filename + '-foreground.png', im_new_foreground)
nparr_labelmap = np.fromstring(labelmap, np.uint8)
labelmapimg = cv2.imdecode(nparr_labelmap, 1)
labelmapimg = cv2.resize(labelmapimg, self.picture_size, interpolation=cv2.INTER_NEAREST)
im_new_labelmapimg = np.where(labelmapimg == 1, 255, labelmapimg)
cv2.imwrite(self.filename + '-labelmap.png', im_new_labelmapimg)
# nparr_scoremap = np.fromstring(scoremap, np.uint8)
# scoremapimg = cv2.imdecode(nparr_scoremap, 1)
# scoremapimg = cv2.resize(scoremapimg, self.picture_size, interpolation=cv2.INTER_NEAREST)
# im_new_scoremapimg = np.where(scoremapimg == 1, 255, scoremapimg)
# cv2.imwrite(self.filename + '-scoremap.png', im_new_scoremapimg)
def get_file_content(filePath):
with open(filePath, 'rb') as fp:
return fp.read()
image = get_file_content('./111.jpg')
# """ 调用手势识别 """
# results = client.gesture(image)
# print(results)
# """ 调用图像分割 """
# options = {}
# options["type"] = 'foreground'
# client.bodySeg(image)
res = client.bodySeg(image)
foreground = base64.b64decode(res['foreground'])
labelmap = base64.b64decode(res['labelmap'])
scoremap = base64.b64decode(res['scoremap'])
nparr_foreground = np.fromstring(foreground, np.uint8)
foregroundimg = cv2.imdecode(nparr_foreground, 1)
foregroundimg = cv2.resize(foregroundimg, (2131, 1438),
interpolation=cv2.INTER_NEAREST)
im_new_foreground = np.where(foregroundimg == 1, 10, foregroundimg)
cv2.imwrite('foreground.png', im_new_foreground)
nparr_labelmap = np.fromstring(labelmap, np.uint8)
labelmapimg = cv2.imdecode(nparr_labelmap, 1)
labelmapimg = cv2.resize(labelmapimg, (2131, 1438),