利用K-NN通过PCA进行图片分类

这是一个小项目，用K-NN通过PCA进行图片分类。

首先惯例，破竹先放出github

数据是用cifar-10的第一个batch的前1000个图片

    K    = int(sys.argv[1])
    D    = int(sys.argv[2])
    N    = int(sys.argv[3])
    file = sys.argv[4]

分别代表：k个邻居，PCA降到d维，0到n是test，n到1000是train

upload数据分好train和test

#extract the data from data_batch_1
def unpickle(file):
    import pickle
    with open(file, 'rb') as fo:
        dict = pickle.load(fo, encoding='latin1')
    return dict

#divide train set and test set
def divide_train_and_test(dict, n):
    data         = dict['data']
    data_test    = data[ : n]
    data_test    = data_test.reshape(n, 3, 32, 32).transpose(0, 2, 3, 1).astype("float")
    data_train   = data[n : 1000]
    data_train   = data_train.reshape(1000 - n, 3, 32, 32).transpose(0, 2, 3, 1).astype("float")
    labels       = dict['labels']
    labels_test  = np.array(labels[ : n])
    labels_train = np.array(labels[n : 1000])
    return data_test, data_train, labels_test, labels_train

然后从RGB变成灰度图