face_recognition报错：operands could not be broadcast together with shapes求助

按教程实现人脸识别，期望返回True/False结果，却触发ValueError，报错信息：operands could not be broadcast together with shapes (1,4) (1,128)。尝试移除np.array、方括号及dtype=object参数后问题仍未解决。

代码 #

import cv2
import numpy as np
import face_recognition
import os

path = 'ImageAttendance'
# 创建文件列表
myList = os.listdir(path)
images = []
classNames = []
print(myList)

# 依次提取每个文件的名称
for cl in myList:
    curImg = cv2.imread(f'{path}/{cl}')
    images.append(curImg)
    classNames.append(os.path.splitext(cl)[0])
print(classNames)
# 计算所有图片的人脸编码并生成列表
def findEncodings(images):
    encodeList = []
    for img in images:
        img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
        encode = face_recognition.face_encodings(img)
        encodeList.append(encode)
    return encodeList

# 生成数据库中所有已知人员的编码列表
encodeListKnown = findEncodings(images)
print(len(encodeListKnown))
print('Encoding complete')

cap = cv2.VideoCapture(0)
while True:
    success, img = cap.read()
    imgS = cv2.resize(img,(0,0), None, 0.25, 0.25)
    imgS = cv2.cvtColor(imgS, cv2.COLOR_BGR2RGB)
    
    
    facesCurFrame = face_recognition.face_locations(imgS)
    encodesCurFrame = face_recognition.face_encodings(imgS, facesCurFrame)
    
    
    for encodeFace,faceLoc in zip(encodesCurFrame, facesCurFrame):
        print('encoding')
        matches = face_recognition.compare_faces(np.array([encodeListKnown], dtype=object),np.array([encodeFace],dtype=object))
        faceDis = face_recognition.face_distance([encodeListKnown], encodeFace)
        print(faceDis)

完整报错信息 #

['face.jpg', 'original.jpg', 'Verc.png', 'test.jpg']
['face', 'original', 'Verc', 'test']
4
Encoding complete
encoding
Traceback (most recent call last):
  File "/home/verc/Документы/Python/FaceRecognition/AttendanceProject.py", line 46, in <module>
    matches = face_recognition.compare_faces(np.array([encodeListKnown], dtype=object),np.array([encodeFace],dtype=object))
  File "/home/verc/.local/lib/python3.10/site-packages/face_recognition/api.py", line 226, in compare_faces
    return list(face_distance(known_face_encodings, face_encoding_to_check) <= tolerance)
  File "/home/verc/.local/lib/python3.10/site-packages/face_recognition/api.py", line 75, in face_distance
    return np.linalg.norm(face_encodings - face_to_compare, axis=1)
ValueError: operands could not be broadcast together with shapes (1,4) (1,128) 
[Finished in 10.8s with exit code 1]
[cmd: ['python3', '-u', '/home/verc/Документы/Python/FaceRecognition/AttendanceProject.py']]
[dir: /home/verc/Документы/Python/FaceRecognition]
[path: /home/verc/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin]

问题核心在于findEncodings函数的实现：face_recognition.face_encodings(img)返回的是包含人脸编码的列表（每张图可能有多个人脸，编码为128维数组），你直接将这个列表追加到encodeList中，导致encodeListKnown变成了嵌套列表（4个元素，每个元素是包含128维数组的列表），而非直接存储128维编码数组的列表。

修改步骤：

1. 修正findEncodings函数，提取每张图的第一个人脸编码（假设每张图只有一个人脸），同时处理无检测到人脸的情况：

def findEncodings(images):
    encodeList = []
    for img in images:
        img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
        encodes = face_recognition.face_encodings(img)
        if encodes:  # 确保检测到人脸
            encodeList.append(encodes[0])
    return encodeList

2. 修改compare_faces和face_distance的调用，不需要手动转成np.array，直接传入处理后的encodeListKnown和encodeFace：

for encodeFace,faceLoc in zip(encodesCurFrame, facesCurFrame):
    print('encoding')
    matches = face_recognition.compare_faces(encodeListKnown, encodeFace)
    faceDis = face_recognition.face_distance(encodeListKnown, encodeFace)
    print(faceDis)

这样修改后，encodeListKnown是由128维编码数组组成的列表，和encodeFace（单个128维数组）的形状匹配，就能正常计算人脸距离和匹配结果了。

内容的提问来源于stack exchange，提问作者blizzadr