函数cvPreprocessCategoricalResponses中的响应数组无效，我为什么要这样做？

Question

我使用Python 2.7和OpenCV构建程序来识别对象。我的数据集是CIFAR-10。我花了两天时间试图理解错误，但我无法找到原因。我的代码是：

import cv2
import cv2.cv as cv
import numpy as np
import os, os.path
import csv

labels = []

with open('labels.csv','rb') as lb:
    for row in lb:
        labels.append(row)

os.chdir('train')
#counts the total number of pictures in the "train" folder
nFiles = len([name for name in os.listdir('.') if os.path.isfile(name)])



extension = '.png'
images = []
imageLabels = []
trainLabels =[]
trainingDescriptors = []
i = 1
sift = cv2.SIFT()
j = 0
BOW = cv2.BOWKMeansTrainer(10)

with open('../trainLabels.csv', 'rb') as labels:
    for row in labels:
        imageLabels.extend(row[1])

while i < 5000:
    #Load all images
    #print "loading file " + str(i) + ".png"
    if i%500 == 0:
        print "Loading file " + str(i)
    image = cv2.imread(str(i)+extension,cv2.CV_LOAD_IMAGE_GRAYSCALE)
    images.append(image.flatten())
    descriptor = np.array([])
    keypoints, descriptor = sift.detectAndCompute(image,None)
    if descriptor != None:
        BOW.add(descriptor)
        trainingDescriptors.extend(descriptor)
    i += 1
    j += 1

dictionary = BOW.cluster()

print "bow dictionary", np.shape(dictionary)

for label in imageLabels:
    if label == 'airplane':
        trainLabels.append(1)
    elif label == 'automobile':
        trainLabels.append(2)
    elif label == 'bird':
        trainLabels.append(3)
    elif label == 'cat':
        trainLabels.append(4)
    elif label == 'deer':
        trainLabels.append(5)
    elif label == 'dog':
        trainLabels.append(6)
    elif label == 'frog':
        trainLabels.append(7)
    elif label == 'horse':
        trainLabels.append(8)
    elif label == 'ship':
        trainLabels.append(9)
    elif label == 'truck':
        trainLabels.append(10)


svm = cv2.SVM()
svm.train(np.array(trainingDescriptors),np.array(trainLabels))

错误在这一行：

svm.train(np.array(trainingDescriptors),np.array(trainLabels))

，错误是：

  

错误：（ - 5）函数cvPreprocessCategoricalResponses中的响应数组无效