Question

我正在尝试运行微表情识别代码，这是链接：https://github.com/bogireddytejareddy/micro-expression-recognition

我遇到一个特定的错误，这是引发错误的代码：

training_list = [a.reshape(-1, 2) for a in training_list]
training_list = numpy.asarray(training_list)
trainingsamples = len(training_list)
traininglabels = numpy.zeros((trainingsamples, ), dtype = int)

traininglabels[0:66] = 0
traininglabels[66:113] = 1
traininglabels[113:156] = 2

traininglabels = np_utils.to_categorical(traininglabels, 3)

training_data = [training_list, traininglabels]
(trainingframes, traininglabels) = (training_data[0], training_data[1])
training_set = numpy.zeros((trainingsamples, 1, image_rows, image_columns, image_depth))

for h in range(trainingsamples):
    training_set[h][0][:][:][:] = trainingframes[h, :, :, :]

training_set = training_set.astype('float32')
training_set -= numpy.mean(training_set)
training_set /= numpy.max(training_set)

错误窗口是：


Using TensorFlow backend.
Traceback (most recent call last):
  File "/Users/Rajat_Sharma/PycharmProjects/untitled1/micro-expression-recognition-master/SMIC/MicroExpSTCNN.py", line 93, in <module>
    training_set[h][0][:][:][:] = trainingframes[h, :, :, :]
IndexError: too many indices for array

Process finished with exit code 1

请帮助。原谅我犯了明显的错误，但我是初学者。

编辑：training_set的形状打印为（7，1，64，64，18），trainingframes的形状打印为（7，）现在，请告诉我代码中要进行哪些更改？许多thanx。

Answer 1

错误非常明显，括号中的索引数比数组的维数多。

检查训练集和训练框架的形状->打印（training_set.shape（））

IndexError：数组，trainingframes [h，：，：，：]的索引过多

1 个答案: