From f9170fcfce4cdd03e9ed38a4adfb48516f378bfd Mon Sep 17 00:00:00 2001 From: tink2123 Date: Sat, 19 Sep 2020 14:40:13 +0800 Subject: [PATCH] polish gen_label --- doc/doc_ch/detection.md | 9 ++++++ doc/doc_ch/recognition.md | 7 +++++ train_data/gen_label.py | 60 +++++++++++++++++++-------------------- 3 files changed, 45 insertions(+), 31 deletions(-) diff --git a/doc/doc_ch/detection.md b/doc/doc_ch/detection.md index 84c90d18..aa320d62 100644 --- a/doc/doc_ch/detection.md +++ b/doc/doc_ch/detection.md @@ -14,6 +14,15 @@ wget -P ./train_data/ https://paddleocr.bj.bcebos.com/dataset/train_icdar2015_l wget -P ./train_data/ https://paddleocr.bj.bcebos.com/dataset/test_icdar2015_label.txt ``` +PaddleOCR 也提供了数据格式转换脚本，可以将官网 label 转换支持的数据格式。 数据转换工具在 `train_data/gen_label.py`, 这里以训练集为例： + +``` +# 将官网下载的标签文件转换为 train_icdar2015_label.txt +python gen_label.py --mode="det" --root_path="icdar_c4_train_imgs/" \ + --input_path="ch4_training_localization_transcription_gt" \ + --output_label="train_icdar2015_label.txt" +``` + 解压数据集和下载标注文件后，PaddleOCR/train_data/ 有两个文件夹和两个文件，分别是： ``` /PaddleOCR/train_data/icdar2015/text_localization/ diff --git a/doc/doc_ch/recognition.md b/doc/doc_ch/recognition.md index c554b9f1..6d6034a6 100644 --- a/doc/doc_ch/recognition.md +++ b/doc/doc_ch/recognition.md @@ -44,6 +44,13 @@ wget -P ./train_data/ic15_data https://paddleocr.bj.bcebos.com/dataset/rec_gt_t wget -P ./train_data/ic15_data https://paddleocr.bj.bcebos.com/dataset/rec_gt_test.txt ``` +PaddleOCR 也提供了数据格式转换脚本，可以将官网 label 转换支持的数据格式。 数据转换工具在 `train_data/gen_label.py`, 这里以训练集为例： + +``` +# 将官网下载的标签文件转换为 rec_gt_label.txt +python gen_label.py --mode="rec" --input_path="{path/of/origin/label}" --output_label="rec_gt_label.txt" +``` + 最终训练集应有如下文件结构： ``` |-train_data diff --git a/train_data/gen_label.py b/train_data/gen_label.py index ae0903b1..de0de2bf 100644 --- a/train_data/gen_label.py +++ b/train_data/gen_label.py @@ -3,38 +3,31 @@ import argparse def gen_rec_label(input_path, out_label): - out_file = open(out_label, 'w') - with open(input_path, 'r') as f: - for line in f.readlines(): - tmp = line.strip('\n').replace(" ", "").split(',') - img_path, label = tmp[0], tmp[1] - label = label.replace("\"", "") - out_file.write(img_path + '\t' + label + '\n') - out_file.close() + with open(out_label, 'w') as out_file: + with open(input_path, 'r') as f: + for line in f.readlines(): + tmp = line.strip('\n').replace(" ", "").split(',') + img_path, label = tmp[0], tmp[1] + label = label.replace("\"", "") + out_file.write(img_path + '\t' + label + '\n') -def gen_det_label(input_dir, out_label): - root_path = "" - if "training" in input_dir: - root_path = "icdar_c4_train_imgs/" - elif "test" in input_dir: - root_path = "ch4_test_images/" - out_file = open(out_label, 'w') - for label_file in os.listdir(input_dir): - img_path = root_path + label_file[3:-4] + ".jpg" - label = [] - with open(os.path.join(input_dir, label_file), 'r') as f: - for line in f.readlines(): - tmp = line.strip("\n\r").replace("\xef\xbb\xbf", "").split(',') - points = tmp[:-2] - s = [] - for i in range(0, len(points), 2): - b = points[i:i + 2] - s.append(b) - result = {"transcription": tmp[-1], "points": s} - label.append(result) - out_file.write(img_path + '\t' + str(label) + '\n') - out_file.close() +def gen_det_label(root_path, input_dir, out_label): + with open(out_label, 'w') as out_file: + for label_file in os.listdir(input_dir): + img_path = root_path + label_file[3:-4] + ".jpg" + label = [] + with open(os.path.join(input_dir, label_file), 'r') as f: + for line in f.readlines(): + tmp = line.strip("\n\r").replace("\xef\xbb\xbf", "").split(',') + points = tmp[:-2] + s = [] + for i in range(0, len(points), 2): + b = points[i:i + 2] + s.append(b) + result = {"transcription": tmp[-1], "points": s} + label.append(result) + out_file.write(img_path + '\t' + str(label) + '\n') if __name__ == "__main__": @@ -44,6 +37,11 @@ if __name__ == "__main__": type=str, default="rec", help='Generate rec_label or det_label, can be set rec or det') + parser.add_argument( + '--root_path', + type=str, + default=".", + help='The root directory of images.Only takes effect when mode=det ') parser.add_argument( '--input_path', type=str, @@ -60,4 +58,4 @@ if __name__ == "__main__": print("Generate rec label") gen_rec_label(args.input_path, args.output_label) elif args.mode == "det": - gen_det_label(args.input_path, args.output_label) + gen_det_label(args.root_path, args.input_path, args.output_label) -- GitLab