添加科大讯了文字检测数据集

README.md

@@ -38,6 +38,7 @@
 | mjsynth                             | http://www.robots.ox.ac.uk/~vgg/data/text/                   | 识别      | 语言: 英文     9,000,000                                     | -                                                            | -                                                            |
 | Synthetic Chinese String  Dataset(360万中文数据集)   | 链接：https://pan.baidu.com/s/1jefn4Jh4jHjQdiWoanjKpQ 提取码：spyi | 识别      | 语言: 混合     300k                                          | -                                                            | -                                                            |
 | 英文识别数据大礼包(https://github.com/clovaai/deep-text-recognition-benchmark) 训练：MJSynth和SynthText  验证：IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE   | 链接：https://pan.baidu.com/s/1KSNLv4EY3zFWHpBYlpFCBQ 提取码：rryk | 识别      | 语言: 英文                                              | -                                                            | -                                                            |
+| 科大讯飞自然场景文字检测挑战赛   | http://challenge.xfyun.cn/topic/info?type=text-detect | 检测      | 语言: 混合                  训练: 5000    验证: 1000                             | {     “gt_1”: [  {“points”: [[x1, y1], [x2, y2], …, [xn,  yn]], "illegibility": false }     | -                                                            | points: x1,y1,x2,y2,x3,y3,x4,y4…xn,yn illegibility: 是否模糊
 
 # 数据生成工具
                                                     

convert/det/check_json.py

@@ -11,7 +11,7 @@
 from convert.utils import show_bbox_on_image, load_gt
 
 if __name__ == '__main__':
-    json_path = r'D:\dataset\icdar2017rctw\detection\train.json'
+    json_path = r'D:\dataset\自然场景文字检测挑战赛初赛数据\验证集\validation_new.json'
     data = load_gt(json_path)
     for img_path, gt in tqdm(data.items()):
         # print(gt['illegibility_list'])

convert/det/iflytek_text_detection.py

@@ -0,0 +1,37 @@
+# -*- coding: utf-8 -*-
+# @Time    : 2020/7/7 10:08
+# @Author  : zhoujun
+import numpy as np
+from tqdm import tqdm
+from convert.utils import load, save
+
+
+def cvt(gt_path, save_path, imgs_folder):
+    gt_dict = {'data_root': imgs_folder}
+    data_list = []
+    ct = load(gt_path)
+
+    for img_id, anns in tqdm(ct.items()):
+        img_name = img_id.replace('gt', 'img') + '.jpg'
+        cur_gt = {'img_name': img_name, 'annotations': []}
+        for ann in anns:
+            cur_line_gt = {'polygon': [], 'text': '', 'illegibility': False, 'language': 'Latin'}
+            chars_gt = [{'polygon': [], 'char': '', 'illegibility': False, 'language': 'Latin'}]
+            cur_line_gt['chars'] = chars_gt
+
+            cur_line_gt['polygon'] = ann['points']
+            cur_line_gt['illegibility'] = ann['illegibility']
+            cur_gt['annotations'].append(cur_line_gt)
+        if len(cur_gt['annotations']) > 0:
+            data_list.append(cur_gt)
+    gt_dict['data_list'] = data_list
+    save(gt_dict, save_path)
+    print(len(gt_dict), len(data_list))
+
+
+if __name__ == '__main__':
+    gt_path = r'D:\dataset\自然场景文字检测挑战赛初赛数据\验证集\validation.json'
+    imgs_folder = r'D:\dataset\自然场景文字检测挑战赛初赛数据\验证集\new_image'
+    save_path = r'D:\dataset\自然场景文字检测挑战赛初赛数据\验证集\validation_new.json'
+    cvt(gt_path, save_path, imgs_folder)
+    # show_coco(gt_path, imgs_folder)

convert/utils.py

@@ -1,14 +1,13 @@
 # -*- coding: utf-8 -*-
 # @Time    : 2020/3/20 19:54
 # @Author  : zhoujun
-import cv2
 import json
 import os
 import glob
 import pathlib
 from natsort import natsorted
 
-__all__ = ['load']
+__all__ = ['load', 'save', 'get_file_list', 'show_bbox_on_image', 'load_gt']
 
 
 def get_file_list(folder_path: str, p_postfix: list = None) -> list:
@@ -85,6 +84,8 @@ def show_bbox_on_image(image, polygons=None, txt=None, color=None, font_path='co
     from PIL import ImageDraw, ImageFont
     image = image.convert('RGB')
     draw = ImageDraw.Draw(image)
+    if len(txt) == 0:
+        txt = None
     if color is None:
         color = (255, 0, 0)
     if txt is not None:
@@ -113,7 +114,7 @@ def load_gt(json_path):
         illegibility_list = []
         language_list = []
         for annotation in gt['annotations']:
-            if len(annotation['polygon']) == 0 or len(annotation['text']) == 0:
+            if len(annotation['polygon']) == 0:
                 continue
             polygons.append(annotation['polygon'])
             texts.append(annotation['text'])