#!/usr/bin/env python # -*- coding: UTF-8 -*- import os, cv2 from pickled import * data_path = './data/train' file_list = './data/train/images.lst' save_path = './bin' testing_data_path = './data/testing' testing_file_list = './data/testing/images.lst' if __name__ == '__main__': #Build training data build_filelist(data_path, file_list) data, label, lst = read_data(file_list, data_path, shape=[32, 32]) pickled(save_path, data, label, lst, bin_num=2) build_meta(data_path, save_path) #Build test data build_filelist(testing_data_path, testing_file_list) data, label, lst = read_data(testing_file_list, testing_data_path, shape=[32, 32]) pickled(save_path, data, label, lst, bin_num=1, mode="test")
# -*- coding: utf-8 -*- import os, cv2 from pickled import * from load_data import * data_path = './data' file_list = './data/images.lst' save_path = './bin' if __name__ == '__main__': data, label, lst = read_data(file_list, data_path, shape=32) pickled(save_path, data, label, lst, bin_num=1)
os.path.dirname(current_path) + os.path.sep + ".") source_image_path = os.path.join( os.path.abspath(os.path.dirname(current_path) + os.path.sep + "../.."), "Source_image") mosaic_image_path = os.path.join( os.path.abspath(os.path.dirname(current_path) + os.path.sep + "../.."), "Mosaic_image") dataset_train_path = os.path.join( os.path.abspath(os.path.dirname(current_path) + os.path.sep + "../.."), "Data_train") dataset_valiate_path = os.path.join( os.path.abspath(os.path.dirname(current_path) + os.path.sep + "../.."), "Data_valiate") #Path of data and datasets image_train_path = source_image_path image_train_record_file = source_image_path + os.path.sep + "data.json" dataset_train_save_path = dataset_train_path dataset_valiate_save_path = dataset_valiate_path # pickled train_dataset data, label, file_name_list = read_data(image_train_record_file, image_train_path, shape=32) pickled(dataset_train_save_path, data, label, file_name_list, bin_num=1, mode="train")
parser = ArgumentParser() parser.add_argument('--folder_path', default="train", help='choose a image folder') parser.add_argument('--mode', default="train", help='train or test') parser.add_argument('--read_image', default="ori_data", help='choose read_data or face_encoding') args = parser.parse_args() data_path = args.folder_path file_list = 'image_list/image_{}_list.txt'.format(data_path) save_path = './bin' mode = args.mode if __name__ == '__main__': if os.path.isfile(file_list): os.remove(file_list) imagelist(data_path, file_list) if args.read_image == 'ori_data': data, label, lst = read_data(file_list, data_path, shape=80) elif args.read_image == 'face_encoding': data, label, lst = face_encoding_read(file_list, data_path) pickled(save_path, data, label, lst, mode, args.read_image, data_path, bin_num=1)
train_image = np.reshape(np.stack(train_image, axis=0), [num_cifar_train, 32*32*3]) train_label = np.reshape(np.array(np.stack(train_label, axis=0)), [num_cifar_train]) fd = os.path.join(target_path, 'test_batch') dict = unpickle(fd) test_image = np.reshape(dict['data'], [num_cifar_test, 32*32*3]) test_label = np.reshape(dict['labels'], [num_cifar_test]) prepare_h5py(train_image, train_label, test_image, test_label, data_dir, [32, 32, 3]) if __name__ == '__main__': #代码未解耦合&判存,每次仅active其中一步 其余注释掉 #1 # resize_pic(image_dir,resized_dir) #train # resize_pic(image_dir_test,resized_dir_test)#test #2. for Lung experiment I re-write the code in Lab's computer and get the .lst files ####Make_pic_list(resized_dir, lst_file_outdir) # train # #3. data, label, lst = read_data(lst_file_outdir, resized_dir, shape=dim) pickled(save_path, data, label, lst_file_outdir, bin_num) # # # 4. args = parser.parse_args() if not os.path.exists(save_path): os.mkdir(save_path) data_process_h5py(save_path) #本文件用于封装数据 最后获得 data.hy 和 id.txt 作为SSGAN网络的输入文件