获取非人脸数据集

实验七/1.get_dataset.py

@@ -2,6 +2,9 @@ import os
 import requests
 import tqdm
 import tarfile
+import pickle
+import numpy as np
+from PIL import Image
 
 # 计算大小
 def get_human_readable_size(size_in_bytes):
@@ -72,6 +75,39 @@ face_label_path='cache/dataset/face/label/'
 face_label_url='http://vis-www.cs.umass.edu/fddb/FDDB-folds.tgz'
 download(face_label_url,face_label_path)
 
+# 下载非人脸数据集
+non_face_path='cache/dataset/non_face/'
+non_face_url='https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz'
+download(non_face_url,non_face_path)
+
 # 解压数据集
 decompress(os.path.join(face_dataset_path,face_dataset_url.split('/')[-1]),face_dataset_path)
-decompress(os.path.join(face_label_path,face_label_url.split('/')[-1]),face_label_path)
+decompress(os.path.join(face_label_path,face_label_url.split('/')[-1]),face_label_path)
+decompress(os.path.join(non_face_path,non_face_url.split('/')[-1]),non_face_path)
+
+# 解码非人脸数据集
+def unpickle(file_path):
+    with open(file_path, 'rb') as file:
+        return pickle.load(file, encoding='bytes')
+
+def save_images(data, labels, directory):
+    os.makedirs(directory, exist_ok=True)
+    for i in tqdm.tqdm(range(len(data))):
+        img = data[i].reshape((3, 32, 32)).transpose((1, 2, 0))
+        img = img.astype(np.uint8)
+        label = str(labels[i])
+        img_filename = os.path.join(directory, f"{label}_{i}.png")
+        img_to_save = Image.fromarray(img)
+        img_to_save.save(img_filename)
+
+print('解码非人脸数据集：')
+cifar10_dir = os.path.join(non_face_path, 'cifar-10-batches-py')
+output_dir = non_face_path
+batch_file = 'data_batch_1'
+file_path = os.path.join(cifar10_dir, batch_file)
+batch_data = unpickle(file_path)
+
+images = batch_data[b'data']
+labels = batch_data[b'labels']
+
+save_images(images, labels, output_dir)