如何利用insightface同时计算多个人脸相似度

1、安装环境：

pip install mxnet,python-opencv,tqdm

2、对齐矫正人脸并裁剪

3、用insightface计算相似度，无需安装insightface等，两个脚本即可：

'''
face_analysis.py
'''from __future__ import division
import mxnet as mx
import numpy as np
import os, cv2
import glob
import zipfile
import hashlib
import requests
from tqdm import tqdmclass FaceRecognition:def __init__(self, name, download, param_file):self.name = nameself.download = downloadself.param_file = param_fileself.image_size = (112, 112)if download:assert param_filedef prepare(self, ctx_id, batch_size=1):if self.param_file:pos = self.param_file.rfind('-')prefix = self.param_file[0:pos]pos2 = self.param_file.rfind('.')epoch = int(self.param_file[pos+1:pos2])sym, arg_params, aux_params = mx.model.load_checkpoint(prefix, epoch)all_layers = sym.get_internals()sym = all_layers['fc1_output']if ctx_id>=0:ctx = mx.gpu(ctx_id)else:ctx = mx.cpu()model = mx.mod.Module(symbol=sym, context=ctx, label_names = None)data_shape = (batch_size,3)+self.image_sizemodel.bind(data_shapes=[('data', data_shape)])model.set_params(arg_params, aux_params)#warmupdata = mx.nd.zeros(shape=data_shape)db = mx.io.DataBatch(data=(data,))model.forward(db, is_train=False)embedding = model.get_outputs()[0].asnumpy()self.model = modelelse:passdef get_embedding(self, data):data = mx.nd.array(data)db = mx.io.DataBatch(data=(data,))self.model.forward(db, is_train=False)embedding = self.model.get_outputs()[0].asnumpy()return embeddingdef compute_sim(self, img1, img2):emb1 = self.get_embedding(img1).flatten()emb2 = self.get_embedding(img2).flatten()from numpy.linalg import normsim = np.dot(emb1, emb2)/(norm(emb1)*norm(emb2))return simdef get_arcface(name, download=True,root='insightface/models', **kwargs):if not download:return FaceRecognition(name, False, None)else:_file = get_model_file("arcface_%s"%name, root=root)return FaceRecognition(name, True, _file)def arcface_r100_v1(**kwargs):return get_arcface("r100_v1", download=True, **kwargs)def check_sha1(filename, sha1_hash):"""Check whether the sha1 hash of the file content matches the expected hash.Parameters----------filename : strPath to the file.sha1_hash : strExpected sha1 hash in hexadecimal digits.Returns-------boolWhether the file content matches the expected hash."""sha1 = hashlib.sha1()with open(filename, 'rb') as f:while True:data = f.read(1048576)if not data:breaksha1.update(data)sha1_file = sha1.hexdigest()l = min(len(sha1_file), len(sha1_hash))return sha1.hexdigest()[0:l] == sha1_hash[0:l]def download(url, path=None, overwrite=False, sha1_hash=None):"""Download an given URLParameters----------url : strURL to downloadpath : str, optionalDestination path to store downloaded file. By default stores to thecurrent directory with same name as in url.overwrite : bool, optionalWhether to overwrite destination file if already exists.sha1_hash : str, optionalExpected sha1 hash in hexadecimal digits. Will ignore existing file when hash is specifiedbut doesn't match.Returns-------strThe file path of the downloaded file."""if path is None:fname = url.split('/')[-1]else:path = os.path.expanduser(path)if os.path.isdir(path):fname = os.path.join(path, url.split('/')[-1])else:fname = pathif overwrite or not os.path.exists(fname) or (sha1_hash and not check_sha1(fname, sha1_hash)):dirname = os.path.dirname(os.path.abspath(os.path.expanduser(fname)))if not os.path.exists(dirname):os.makedirs(dirname)print('Downloading %s from %s...'%(fname, url))r = requests.get(url, stream=True)if r.status_code != 200:raise RuntimeError("Failed downloading url %s"%url)total_length = r.headers.get('content-length')with open(fname, 'wb') as f:if total_length is None: # no content length headerfor chunk in r.iter_content(chunk_size=1024):if chunk: # filter out keep-alive new chunksf.write(chunk)else:total_length = int(total_length)for chunk in tqdm(r.iter_content(chunk_size=1024),total=int(total_length / 1024. + 0.5),unit='KB', unit_scale=False, dynamic_ncols=True):f.write(chunk)if sha1_hash and not check_sha1(fname, sha1_hash):raise UserWarning('File {} is downloaded but the content hash does not match. ' \'The repo may be outdated or download may be incomplete. ' \'If the "repo_url" is overridden, consider switching to ' \'the default repo.'.format(fname))return fnamedef find_params_file(dir_path):if not os.path.exists(dir_path):return Nonepaths = glob.glob("%s/*.params"%dir_path)if len(paths)==0:return Nonepaths = sorted(paths)return paths[-1]def get_model_file(name, root=os.path.join('insightface', 'models')):r"""Return location for the pretrained on local file system.This function will download from online model zoo when model cannot be found or has mismatch.The root directory will be created if it doesn't exist.Parameters----------name : strName of the model.root : str, default '~/.mxnet/models'Location for keeping the model parameters.Returns-------file_pathPath to the requested pretrained model file."""_model_sha1 = {name: checksum for checksum, name in [('95be21b58e29e9c1237f229dae534bd854009ce0', 'arcface_r100_v1'),('', 'arcface_mfn_v1'),('39fd1e087a2a2ed70a154ac01fecaa86c315d01b', 'retinaface_r50_v1'),('2c9de8116d1f448fd1d4661f90308faae34c990a', 'retinaface_mnet025_v1'),('0db1d07921d005e6c9a5b38e059452fc5645e5a4', 'retinaface_mnet025_v2'),('7dd8111652b7aac2490c5dcddeb268e53ac643e6', 'genderage_v1'),]}base_repo_url = 'http://insightface.ai/files/'_url_format = '{repo_url}models/{file_name}.zip'file_name = nameroot = os.path.expanduser(root)dir_path = os.path.join(root, name)file_path = find_params_file(dir_path)#file_path = os.path.join(root, file_name + '.params')sha1_hash = _model_sha1[name]if file_path is not None:if check_sha1(file_path, sha1_hash):return file_pathelse:print('Mismatch in the content of model file detected. Downloading again.')else:print('Model file is not found. Downloading.')if not os.path.exists(root):os.makedirs(root)if not os.path.exists(dir_path):os.makedirs(dir_path)zip_file_path = os.path.join(root, file_name + '.zip')repo_url = base_repo_urlif repo_url[-1] != '/':repo_url = repo_url + '/'download(_url_format.format(repo_url=repo_url, file_name=file_name),path=zip_file_path,overwrite=True)with zipfile.ZipFile(zip_file_path) as zf:zf.extractall(dir_path)os.remove(zip_file_path)file_path = find_params_file(dir_path)if check_sha1(file_path, sha1_hash):return file_pathelse:raise ValueError('Downloaded file has different hash. Please try again.')

调用face_analysis.py即可多张图片同时inference求相似度，也可以将结果放到pytorch中调用：

from face_analysis import *
import cv2
import numpy as np
import torchclass FaceAnalysis:def __init__(self, ctx_id=-1, batch_size=1, rec_name='arcface_r100_v1'):if rec_name is not None:self.rec_model = arcface_r100_v1()else:self.rec_model = Noneif self.rec_model is not None:self.rec_model.prepare(ctx_id=ctx_id, batch_size=batch_size)def get(self, data):# if self.rec_model is not None:#     self.rec_model.prepare(ctx_id, batch_size=1)embedding = Noneif self.rec_model is not None:embedding = self.rec_model.get_embedding(data)return embeddinginsface_model = FaceAnalysis(ctx_id=-1)#cpu:-1,gpu:1img = cv2.imread('face.png')img = cv2.resize(img[19:237, 19:237], (112, 112))
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
data = np.transpose(img, (2,0,1))
data = np.expand_dims(data, axis=0)img2 = cv2.imread('girl.png')
img2 = cv2.resize(img2[19:237, 19:237], (112, 112))
img2 = cv2.cvtColor(img2, cv2.COLOR_BGR2RGB)
data2 = np.transpose(img2, (2,0,1))
data2 = np.expand_dims(data2, axis=0)a= []
a.append(data)
a.append(data2)
datas = np.vstack(a)
batch_size = 2
emb = insface_model.get(datas)
print(emb.shape)
emb1 = emb[0].flatten()
emb2 = emb[1].flatten()
from numpy.linalg import norm
sim = np.dot(emb1, emb2)/(norm(emb1)*norm(emb2))
print(sim)
# embedding_norm = norm(embedding)
# normed_embedding = embedding / embedding_norm
emb = =torch.from_numpy(emb)

如何利用insightface同时计算多个人脸相似度相关推荐

CV之FR之MTCNN：基于TF框架利用MTCNN算法检测并对齐人脸图像进(人脸识别/人脸相似度)而得出人脸特征向量从而计算两张人脸图片距离案例应用之详细攻略
CV之FR之MTCNN:基于TF框架利用MTCNN算法检测并对齐人脸图像进(人脸识别/人脸相似度)而得出人脸特征向量从而计算两张人脸图片距离案例应用之详细攻略目录基于TF框架利用MTCNN算法检测 ...
python 摄像头标定_python 3利用Dlib 19.7实现摄像头人脸检测特征点标定
Python 3 利用 Dlib 19.7 实现摄像头人脸检测特征点标定 0.引言利用python开发,借助Dlib库捕获摄像头中的人脸,进行实时特征点标定: 图1 工程效果示例(gif) 图2 工 ...
利用openssl来计算sha1, sha224, sha256, sha384, sha512
转载:http://blog.csdn.net/stpeace/article/details/42371079 利用openssl来计算sha1, sha224, sha256, sha384, s ...
利用scipy包计算表格线的峰值,还原表格得到表格结构
1. 利用scipy包计算表格线的峰值 import cv2 import numpy as np from scipy.signal import find_peaks, peak_widthsde ...
利用MSFM算法计算程函方程的走时CT胖射线反演法
利用MSFM算法计算程函方程的走时CT胖射线反演法基于程函方程的弹性波走时CT胖射线反演方法,是在反演迭代过程中,利用程函方程计算走时场,并得到接收点处的走时数据,用这个数据代替正演数据.每次迭代完 ...
我的第一个MPI程序:利用矩形规则计算pi
利用矩形规则计算pi #include<mpi.h> #include<iostream> using namespace std; const int INTERVALS ...
怎么利用计算机公式计算完成比例,excel表格数据计算所占比例公式的使用教程...
Excel中经常需要使用到公式计算所占比例,所占比例具体该如何利用函数进行计算呢?接下来是学习啦小编为大家带来的excel表格数据计算所占比例公式的使用教程,供大家参考. excel表格数据计算所占比 ...
python计算圆周率近似值_怎样利用python来计算出圆周率π-百度经验
怎样利用python来计算出圆周率π? 方法/步骤 1 圆周率π是一个无理数,没有任何一个精确公式可以计算出π值,π的计算只能采用近似算法. 我们利用蒙特卡洛方法来求得π的近似值. 蒙特卡洛方法是利用 ...
【JavaBigDecimal练习】利用BigDecimal精确计算欧拉数
利用BigDecimal精确计算欧拉数小数点后面1000位相信大家对于泰勒展开已经稔熟于心,接下来我们就利用下列的泰勒展开: e x = 1 + 1 1 ! x + 1 2 ! x 2 + 1 3 ...
计算机函数if80到90分怎么弄,用EXCEL工作表怎样利用IF函数计算分数等级
用EXCEL工作表怎样利用IF函数计算分数等级 01 打开保存有分数的表格,如图所示.这里假设分数设为 A:90分以上(含90) B:80~89 C:70~79 D:60~69 E:60以下(不含60 ...

如何利用insightface同时计算多个人脸相似度

如何利用insightface同时计算多个人脸相似度相关推荐

最新文章

热门文章