add IC15 官方数据集

aliocr_IC15_convert.py

@@ -1,5 +1,565 @@
 
-# opencv-python==4.6.0.66
+# pip install numpy==1.26.4 opencv-python==4.6.0.66
+
+
+"""
+
+给 PaddleOCR 用，前面是坐标和图片都变换；这里图像不变，坐标不变 
+
+
+将阿里OCR 的识别结果（图片和标注）转换成 icdar2015 格式 (注意：它的文本是含 utf8 bom 的)
+
+给 mmocr 训练用。格式是 icdar2015 的格式，文件夹的组织方式是按照 mmocr 的要求创建的
+
+"""
+
+
+"""
+
+! unzip ./GD500.zip -d DB/datasets
+
+icdar2015 文本检测数据集
+标注格式: x1,y1,x2,y2,x3,y3,x4,y4,text
+
+其中, x1,y1为左上角坐标,x2,y2为右上角坐标,x3,y3为右下角坐标,x4,y4为左下角坐标。 
+
+### 表示text难以辨认。
+"""
+
+
+
+
+import random
+from pathlib import Path
+import os
+import glob
+import base64
+from importlib.resources import path
+import math
+import numpy as np
+import cv2
+import json
+import decimal
+import datetime
+from pickletools import uint8
+class DecimalEncoder(json.JSONEncoder):
+    def default(self, o):
+        if isinstance(o, decimal.Decimal):
+            return float(o)
+        elif isinstance(o, datetime.datetime):
+            return str(o)
+        super(DecimalEncoder, self).default(o)
+
+
+def save_json(filename, dics):
+    with open(filename, 'w', encoding='utf-8') as fp:
+        json.dump(dics, fp, indent=4, cls=DecimalEncoder, ensure_ascii=False)
+        fp.close()
+
+
+def load_json(filename):
+    with open(filename, encoding='utf-8') as fp:
+        js = json.load(fp)
+        fp.close()
+        return js
+
+# convert string to json
+
+
+def parse(s):
+    return json.loads(s, strict=False)
+
+# convert dict to string
+
+
+def string(d):
+    return json.dumps(d, cls=DecimalEncoder, ensure_ascii=False)
+
+
+def transform(points, M):
+    # points 算出四个点变换后移动到哪里了
+    # points = np.array([[word_x,  word_y],              # 左上
+    #                    [word_x + word_width, word_y],                 # 右上
+    #                    [word_x + word_width, word_y + word_height],  # 右下
+    #                    [word_x, word_y + word_height],  # 左下
+    #                    ])
+    # add ones
+    ones = np.ones(shape=(len(points), 1))
+
+    points_ones = np.hstack([points, ones])
+
+    # transform points
+    transformed_points = M.dot(points_ones.T).T
+
+    transformed_points_int = np.round(
+        transformed_points, decimals=0).astype(np.int32)  # 批量四舍五入
+    
+    return transformed_points_int
+
+
+def cutPoly(img, pts):
+    # img = cv2.imdecode(np.fromfile('./t.png', dtype=np.uint8), -1)
+    # pts = np.array([[10,150],[150,100],[300,150],[350,100],[310,20],[35,10]])
+
+    ## (1) Crop the bounding rect
+    rect = cv2.boundingRect(pts)
+    x,y,w,h = rect
+    croped = img[y:y+h, x:x+w].copy()
+
+    ## (2) make mask
+    pts = pts - pts.min(axis=0)
+
+    mask = np.zeros(croped.shape[:2], np.uint8)
+    cv2.drawContours(mask, [pts], -1, (255, 255, 255), -1, cv2.LINE_AA)
+
+    ## (3) do bit-op
+    dst = cv2.bitwise_and(croped, croped, mask=mask)
+
+    ## (4) add the white background
+    bg = np.ones_like(croped, np.uint8)*255
+    cv2.bitwise_not(bg,bg, mask=mask)
+    dst2 = bg+ dst
+
+
+    # cv2.imwrite("croped.png", croped)
+    # cv2.imwrite("mask.png", mask)
+    # cv2.imwrite("dst.png", dst)
+    # cv2.imwrite("dst2.png", dst2)
+
+    return dst2
+
+
+
+if __name__ == "__main__":
+
+    # 验证原版的文本标记框
+    im = './train_images/img_1.jpg'
+    gt = './train_gts/gt_img_1.txt'
+
+    # 验证自已生成的标记框
+    # im = './icdar2015_aliocr/imgs/training/img_1.jpg'
+    # gt = './icdar2015_aliocr/annotations/training/gt_img_1.txt'
+
+    if os.path.exists(gt):
+
+        items = []
+        reader = open(gt, 'r', encoding='utf-8-sig').readlines()
+        for line in reader:
+            item = {}
+            parts = line.strip().split(',')
+            label = parts[-1]
+            if 'TD' in gt and label == '1':
+                label = '###'
+            line = [i.strip('\ufeff').strip('\xef\xbb\xbf') for i in parts]
+            if 'icdar' in gt:
+                poly = np.array(list(map(float, line[:8]))).reshape(
+                    (-1, 2)).tolist()
+            else:
+                num_points = math.floor((len(line) - 1) / 2) * 2
+                poly = np.array(list(map(float, line[:num_points]))).reshape(
+                    (-1, 2)).tolist()
+            item['poly'] = poly
+            item['text'] = label
+            # 多边形是用一个个的点表示的，起点连接第二个点，第二个连接第三个 ... 最后一点连接起点，构成一个闭合的区域
+            item['points'] = poly
+            # 此标记表示文字模糊不可辨认，文本框的标记是不可靠的
+            item['ignore'] = True if label == '###' else False
+            items.append(item)
+
+        img = cv2.imdecode(np.fromfile(im, dtype=np.uint8), -1)
+        # DBNet 原版代码只能处理彩图，所以统一处理成彩图
+        img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
+
+        for i in range(len(items)):
+            poly = items[i]['poly']
+            poly = np.array(poly)
+            poly = poly.astype(np.int32)
+
+            #cv2.fillPoly(img, pts=[ poly ], color=(0, 0, 255))
+
+            b = random.randint(0, 255)  # 用来生成[a,b]之间的随意整数，包括两个边界值。
+            g = random.randint(0, 255)
+            r = random.randint(0, 255)
+
+            # 只画线，不填充  # 就是画线，从起点连到第二个点 ... 最后一个点连到第一个点
+            cv2.polylines(img, [poly], isClosed=True,
+                          color=(b, g, r), thickness=1)
+
+        #cv2.imwrite("poly.jpg", img)
+
+        # cv2.imshow("poly", img)
+        # cv2.waitKey()
+
+    # 开始转换
+
+    out_dir = 'icdar2015_aliocr'
+
+    # https://help.aliyun.com/document_detail/294540.html 阿里云ocr结果字段定义
+    # prism-wordsInfo 里的 angle 文字块的角度，这个角度只影响width和height，当角度为-90、90、-270、270，width和height的值需要自行互换
+
+    dir_json = './data/json'  # '/yingedu/www/ocr_server/data/json'
+    dir_img = './data/img'  # '/yingedu/www/ocr_server/data/img'
+
+    train_list = []
+    train_list_txt_path = os.path.join(out_dir, 'train_list.txt')
+
+    g_count = 1
+
+    json_paths = glob.glob('{}/*.json'.format(dir_json), recursive=True)
+
+    for json_path in json_paths:
+
+        base = Path(json_path).stem
+
+        img_path = os.path.join(dir_img, '{}.txt'.format(base))
+
+        if not os.path.exists(img_path):  # 没有相应的图片，可能被删除了
+            continue
+
+        jsn = load_json(json_path)
+
+        with open(img_path, "r", encoding="utf-8") as fp:
+            imgdata = fp.read()
+            imgdata = base64.b64decode(imgdata)
+            imgdata = np.frombuffer(imgdata, np.uint8)
+            img = cv2.imdecode(imgdata, cv2.IMREAD_UNCHANGED)
+
+            # cv2.imshow('img', img)
+            # cv2.waitKey(0)
+
+        if len(img.shape) != 3:  # 转彩图
+            img_color = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
+            img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)  # DBNet 原版只能处理彩图，这里转一下
+
+        else:
+            img_color = img.copy()
+
+        img_color_origin = img_color.copy()
+        img_color_origin2 = img_color.copy()
+
+
+        # 生成1000 张一模一样的图
+        for i in range(1, 1000+1):
+
+            num_img = i
+
+            img_name = "img_{}.jpg".format(num_img)
+            gt_name = "gt_img_{}.txt".format(num_img)
+
+            gt_txt_list = []
+
+            train_list.append(img_name)
+            # num_img += 1
+
+            img_path = os.path.join(out_dir, 'imgs', 'training', img_name)
+            img_gt_path = os.path.join(
+                out_dir, 'annotations', 'training', gt_name)
+
+            cv2.imwrite(img_path, img)
+
+            wordsInfo = jsn['prism_wordsInfo']
+            for j in range(len(wordsInfo)):
+                jo = wordsInfo[j]
+                word = jo["word"]
+                # prism-wordsInfo 里的 angle 文字块的角度，这个角度只影响width和height，当角度为-90、90、-270、270，width和height的值需要自行互换
+                angle = jo['angle']
+                
+                img_color = img_color_origin.copy()
+
+                word_x = jo['x']
+                word_y = jo['y']
+                word_width = jo['width']
+                word_height = jo['height']
+
+                if abs(angle) == 90 or abs(angle) == 270:
+                    word_width = jo['height']
+                    word_height = jo['width']
+                elif angle != 0:
+
+                    # 变换前画出绿框，方便追踪点的前后变化
+                    # img_color = cv2.rectangle(img_color, (word_x, word_y), (
+                    #     word_x + word_width, word_y + word_height), (0, 255, 0), 2)  # 矩形的左上角, 矩形的右下角
+
+                    # cv2.imshow("green", img_color)
+                    # cv2.waitKey(0)
+
+                    # 变换前的多边形蓝框
+                    points = np.array([
+                        [word_x,  word_y],                             # 左上
+                        [word_x + word_width, word_y],                 # 右上
+                        [word_x + word_width, word_y + word_height],  # 右下
+                        [word_x, word_y + word_height],                # 左下
+                    ])
+
+                    # # cv2.fillPoly(img_color, pts=[points], color=(255, 0, 0)) # 填充
+                    # cv2.polylines(img_color, [points], isClosed=True, color=(
+                    #     255, 0, 0), thickness=1)  # 只画线，不填充
+
+                    # cv2.imshow("polys", img_color)
+                    # cv2.waitKey(0)
+
+                    # 获取图像的维度，并计算中心
+                    (h, w) = img_color.shape[:2]
+                    (cX, cY) = (w // 2, h // 2)
+
+                    # - (cX,cY): 旋转的中心���坐标
+                    # - 180: 旋转的度数，正度数表示逆时针旋转，而负度数表示顺时针旋转。
+                    # - 1.0：旋转后图像的大小，1.0原图，2.0变成原来的2倍，0.5变成原来的0.5倍
+                    # 1° = π/180弧度   1 弧度 =  180 / 3.1415926   // 0.0190033 是Mathematica 算出来的弧度，先转换成角度  // -0.0190033 * (180 / 3.1415926)
+                    M = cv2.getRotationMatrix2D((cX, cY), angle, 1.0)
+                    img_color = cv2.warpAffine(img_color, M, (w, h))
+                    img_color_transform = img_color.copy()
+
+                    # cv2.imshow("after trans", img_color)
+                    # cv2.waitKey(0)
+
+                    # https://docs.opencv.org/2.4/doc/tutorials/imgproc/imgtrans/warp_affine/warp_affine.html  # 原理
+                    # https://stackoverflow.com/questions/30327659/how-can-i-remap-a-point-after-an-image-rotation # How can I remap a point after an image rotation?
+                    # 如何得到移动后的坐标点
+
+                    # points 算出四个点变换后移动到哪里了
+                    points = np.array([[word_x,  word_y],              # 左上
+                                       # 右上
+                                       [word_x + word_width, word_y],
+                                       [word_x + word_width, word_y + \
+                                           word_height],  # 右下
+                                       [word_x, word_y + word_height],  # 左下
+                                       ])
+                    # add ones
+                    ones = np.ones(shape=(len(points), 1))
+
+                    points_ones = np.hstack([points, ones])
+
+                    # transform points
+                    transformed_points = M.dot(points_ones.T).T
+
+                    transformed_points_int = np.round(
+                        transformed_points, decimals=0).astype(np.int32)  # 批量四舍五入
+
+                    cv2.polylines(img_color, [transformed_points_int], isClosed=True, color=(
+                        0, 0, 255), thickness=2)  # 画转换后的点
+
+
+                    cv2.polylines(img_color_origin, [points], isClosed=True, color=(
+                        random.randint(0, 255), random.randint(0, 255), random.randint(0, 255)), thickness=2)  # 画转换前的点
+
+                    
+
+                    # cv2.imshow("orgin", img_color_origin)
+                    # cv2.waitKey(0)
+
+
+
+
+                # 四个角的位置 # 左上、右上、右下、左下，当NeedRotate为true时，如果最外层的angle不为0，需要按照angle矫正图片后，坐标才准确
+                pos = jo["pos"]
+                x = int(pos[0]["x"])  # 左上
+                y = int(pos[0]["y"])
+
+                x2 = int(pos[2]["x"])  # 右下
+                y2 = int(pos[2]["y"])
+
+                lu = [pos[0]['x'], pos[0]['y']]  # left up  四个角顺时针方向数
+                ru = [pos[1]['x'], pos[1]['y']]
+                rd = [pos[2]['x'], pos[2]['y']]
+                ld = [pos[3]['x'], pos[3]['y']]
+
+                # 生成 icdar2015 格式的人工标记训练数据（用于训练 mmocr）
+                #gt_txt_list.append( "{},{},{},{},{},{},{},{},{}".format(lu[0], lu[1], ru[0], ru[1], rd[0], rd[1], ld[0], ld[1], word) )
+
+                # 绘制矩形
+                start_point = (x, y)  # 矩形的左上角
+
+                end_point = (x2, y2)  # 矩形的右下角
+
+                color = (0, 0, 255)  # BGR
+
+                thickness = 2
+
+                # 逐行画框
+                # img_color = cv2.rectangle(img_color, start_point, end_point, color, thickness)
+                # cv2.imshow("box", img_color)
+
+                # cv2.waitKey(0)
+
+                points = [ lu, ru, rd, ld ]
+
+
+
+                points0 = np.array([[word_x,  word_y],              # 左上
+                                       # 右上
+                                       [word_x + word_width, word_y],
+                                       [word_x + word_width, word_y + \
+                                           word_height],  # 右下
+                                       [word_x, word_y + word_height],  # 左下
+                                       ])
+                points1 = np.array( [ lu, ru, rd, ld ] )
+
+
+                if not (abs(angle) == 90 or abs(angle) == 270) and angle != 0:
+                   points = transform( points, M )
+                else:
+                   points = np.array(points)
+
+                   ps3 = np.array(                 
+                        [
+                            [min( points[0][0], points1[0][0] ), min( points[0][1], points1[0][1] )], # 左上(取最两者中最小的)
+
+                            [max( points[1][0], points1[1][0] ), min( points[1][1], points1[1][1] )], # 右上
+
+                            [max( points[2][0], points1[2][0] ), max( points[2][1], points1[2][1] )], # 右下
+
+                            [min( points[3][0], points1[3][0] ), max( points[3][1], points1[3][1] )]  # 左下
+                        ]
+                   )
+
+                   img_cuted = cutPoly(img, ps3)
+                   cv2.imwrite(f'./tmp/{g_count}.jpg', img_cuted)
+                   with open(f'./tmp/{g_count}.txt', 'w', encoding='utf-8') as f:
+	                    f.write(word)
+                #    g_count += 1
+
+                # cv2.polylines(img_color, [points], isClosed=True, color=(   # 多边形，框得比较全
+                #     100, 0, 255), thickness=2)  # 只画线，不填充
+
+
+                # cv2.polylines(img_color_origin, [ points1 ], isClosed=True, color=(
+                #     random.randint(0, 255), random.randint(0, 255), random.randint(0, 255)), thickness=2)  # 画转换前的点
+
+                # cv2.imshow("orgin", img_color_origin)
+                # cv2.waitKey(0)
+
+                # # cv2.imshow("box", img_color)
+                # # cv2.waitKey(0)
+
+                # # img_color = cv2.rectangle(img_color, points[0], points[2], color, thickness)  # 正常矩形，框不完全
+                # # cv2.imshow("box", img_color)
+
+                # # cv2.waitKey(0)
+
+
+
+
+
+                # if not (abs(angle) == 90 or abs(angle) == 270) and angle != 0:
+
+                #     t = word
+                #     ps = np.array(                        
+                #         [
+                #             [min( transformed_points_int[0][0], points[0][0] ), min( transformed_points_int[0][1], points[0][1] )], # 左上(取最两者中最小的)
+
+                #             [max( transformed_points_int[1][0], points[1][0] ), min( transformed_points_int[1][1], points[1][1] )], # 右上
+
+                #             [max( transformed_points_int[2][0], points[2][0] ), max( transformed_points_int[2][1], points[2][1] )], # 右下
+
+                #             [min( transformed_points_int[3][0], points[3][0] ), max( transformed_points_int[3][1], points[3][1] )]  # 左下
+                #         ]
+                #     )
+
+
+                #     ps2 = np.array(                 
+                #         [
+                #             [min( points0[0][0], points1[0][0] ), min( points0[0][1], points1[0][1] )], # 左上(取最两者中最小的)
+
+                #             [max( points0[1][0], points1[1][0] ), min( points0[1][1], points1[1][1] )], # 右上
+
+                #             [max( points0[2][0], points1[2][0] ), max( points0[2][1], points1[2][1] )], # 右下
+
+                #             [min( points0[3][0], points1[3][0] ), max( points0[3][1], points1[3][1] )]  # 左下
+                #         ]
+                #     )
+
+                #     # img_cuted = cutPoly(img_color_transform, ps)
+                #     # cv2.imwrite(f'./tmp/{g_count}.jpg', img_cuted)
+
+                #     # with open(f'./tmp/{g_count}.txt', 'w', encoding='utf-8') as f:
+	            #     #     f.write(word)
+                    
+                #     # g_count += 1
+
+                #     cv2.polylines(img_color, [ ps ], isClosed=True, color=(
+                #         255, 0, 0), thickness=2)  # 只画线，不填充
+
+                #     cv2.polylines(img_color_origin, [ ps2 ], isClosed=True, color=(
+                #         random.randint(0, 255), random.randint(0, 255), random.randint(0, 255)), thickness=2)  # 只画线，不填充
+                        
+                #     cv2.imshow("orgin", img_color_origin)
+                #     cv2.waitKey(0)
+
+                #     img_cuted = cutPoly(img, ps2)
+                #     cv2.imwrite(f'./tmp/{g_count}.jpg', img_cuted)
+
+                #     with open(f'./tmp/{g_count}.txt', 'w', encoding='utf-8') as f:
+	            #         f.write(word)
+
+            #         g_count += 1
+
+
+            #         # cv2.imshow("box", img_color)
+
+            #         # cv2.waitKey(0)
+
+            #     lastx_mini = 0  # 下一个字符x 坐标的下界（肯定不小于这个值）
+            #     prew = 0  # 上一个字符的宽度
+            #     words = ""
+            #     charInfo = jo["charInfo"]
+
+            #     min_cx = 9999   # 最小左上角
+            #     min_cy = 9999
+
+            #     max_cxcw = -1   # 最大右下角
+            #     max_cych = -1
+
+            #     for i in range(len(charInfo)):
+            #         joc = charInfo[i]
+            #         c = joc["word"]
+            #         cx = int(joc["x"])
+            #         cy = int(joc["y"])
+            #         cw = int(joc["w"])
+            #         ch = int(joc["h"])
+
+            #         if cx < min_cx:
+            #             min_cx = cx
+            #         if cy < min_cy:
+            #             min_cy = cy
+
+            #         if cx + cw > max_cxcw:
+            #             max_cxcw = cx + cw
+
+            #         if cy + ch > max_cych:
+            #             max_cych = cy + ch
+
+            #         # 绘制矩形
+            #         start_point = (cx, cy)  # 矩形的左上角
+
+            #         end_point = (cx + cw, cy + ch)  # 矩形的右下角
+
+            #         color = (0, 0, 255)  # BGR
+
+            #         thickness = 2
+
+            #         # 逐字画框
+            #         # img_color = cv2.rectangle(
+            #         #     img_color, start_point, end_point, color, thickness)
+            #         # cv2.imshow("box", img_color)
+            #         # cv2.waitKey(0)
+
+            #     # 这个框更准一些
+            #     # img_color = cv2.rectangle(
+            #     #     img_color, (min_cx, min_cy), (max_cxcw, max_cych), (0, 255, 0), thickness)
+            #     # cv2.imshow("box", img_color)
+            #     # cv2.waitKey(0)
+
+            #     # fix me: 如果上面的行框的左边要比这里更左，那就以行框的左边为准
+            #     # 因为发现单个字的框会有漏字的现想
+
+            #     gt_txt_list.append("{},{},{},{},{},{},{},{},{}".format(
+            #         min_cx, min_cy, max_cxcw, min_cy, max_cxcw, max_cych, min_cx, max_cych, word))
+
+            # gt_txt = '\n'.join(gt_txt_list)
+
+            # with open(img_gt_path, "w", encoding='utf-8-sig') as fp:
+            #     fp.write(gt_txt)
 
 
 

datasets/icdar2015/test_gts/gt_img_1.txt

@@ -0,0 +1,8 @@
+933,255,954,255,956,277,936,277,###
+172,323,195,324,195,339,177,339,###
+83,270,118,271,115,294,88,291,###
+940,310,962,310,962,320,940,320,###
+946,356,976,351,978,368,950,374,###
+940,322,962,322,964,333,943,334,###
+128,344,210,342,206,361,128,362,###
+312,303,360,303,360,312,312,312,###

datasets/icdar2015/test_gts/gt_img_10.txt

@@ -0,0 +1,10 @@
+27,17,103,22,106,47,30,45,Please
+107,20,159,26,159,48,109,47,lower
+161,26,198,27,199,51,163,51,your
+201,28,251,31,251,48,201,46,volume
+35,52,97,51,100,76,39,79,when
+101,55,140,53,143,80,103,81,you
+141,55,181,53,183,77,144,79,pass
+182,51,205,52,205,76,185,77,###
+41,83,148,77,151,103,45,113,residential
+152,82,198,80,199,99,153,101,areas

datasets/icdar2015/test_gts/gt_img_100.txt

@@ -0,0 +1,12 @@
+454,115,477,115,474,130,457,132,###
+476,114,524,112,525,132,474,130,diverse
+560,83,630,85,636,113,562,113,tastes
+474,133,567,128,568,142,475,141,###
+523,110,580,115,580,127,525,128,flavours
+515,85,558,83,558,111,517,111,the
+408,82,518,80,520,113,410,114,Refishing
+660,82,717,82,717,101,661,98,DINING
+935,134,963,135,962,152,934,151,###
+657,127,695,127,695,135,657,135,###
+487,182,528,182,528,194,487,194,DINING
+483,194,538,195,538,208,483,207,###

datasets/icdar2015/test_gts/gt_img_101.txt

@@ -0,0 +1,3 @@
+831,115,1003,72,1012,171,835,195,SALE
+410,187,451,192,447,207,406,203,MARC
+451,191,485,194,484,208,450,205,###

datasets/icdar2015/test_gts/gt_img_102.txt

@@ -0,0 +1,4 @@
+615,246,698,251,694,281,615,272,ST.MARC
+698,252,742,261,740,283,697,278,CAFE
+865,223,898,213,901,233,867,243,###
+902,212,945,198,946,217,903,231,###

datasets/icdar2015/test_gts/gt_img_103.txt

@@ -0,0 +1,25 @@
+1058,377,1144,381,1147,412,1062,408,###
+1062,405,1152,406,1155,438,1065,437,###
+23,327,110,325,111,346,24,348,###
+736,266,781,262,783,276,738,281,###
+345,340,428,334,430,364,346,370,###
+750,203,788,202,791,223,752,224,EXIT
+875,232,948,222,948,238,875,248,###
+131,260,404,250,401,325,130,325,ANGKOK
+1024,217,1111,195,1115,226,1028,248,###
+1033,244,1115,228,1125,291,1043,306,KOK
+1038,302,1124,292,1130,317,1044,327,ABUR
+1062,325,1133,320,1135,355,1064,361,###
+1068,355,1141,354,1143,381,1071,382,###
+941,248,1002,233,1006,260,945,275,###
+946,271,967,265,973,315,952,321,###
+960,337,1011,334,1012,364,961,367,###
+958,361,1030,362,1032,386,961,385,PATH
+970,383,1033,383,1034,410,971,410,###
+968,405,1041,406,1040,432,967,431,CHAT
+954,320,978,317,978,341,954,343,###
+738,291,757,288,759,302,740,305,###
+757,290,790,286,790,299,757,303,###
+875,249,922,244,922,259,875,264,###
+802,281,839,279,839,289,802,292,###
+737,279,783,276,782,286,737,289,###

datasets/icdar2015/test_gts/gt_img_104.txt

@@ -0,0 +1,9 @@
+53,106,152,128,151,200,55,197,END
+51,210,160,201,163,326,61,352,###
+78,354,157,335,161,392,84,417,MILE
+67,432,173,388,177,442,74,495,SPEED
+651,108,666,106,666,123,651,125,###
+431,214,475,214,475,223,431,223,###
+1148,207,1180,204,1180,220,1149,223,gels
+1150,188,1166,187,1167,200,1150,201,###
+110,479,151,458,152,475,112,497,###

datasets/icdar2015/test_gts/gt_img_105.txt

@@ -0,0 +1 @@
+0,6,352,94,357,172,1,77,aigonLotus

datasets/icdar2015/test_gts/gt_img_106.txt

@@ -0,0 +1,24 @@
+447,560,475,555,474,567,446,572,###
+892,352,948,376,946,398,892,377,Marina
+911,388,948,402,944,425,910,411,MRT
+952,408,1007,436,1001,454,950,431,###
+1004,434,1046,453,1045,468,1004,452,Station
+950,375,985,391,983,411,948,399,Link
+318,563,434,543,434,555,318,577,MARINA
+314,575,428,554,427,577,307,597,LINK
+1241,38,1255,38,1254,63,1240,63,###
+396,121,447,123,448,136,397,134,PARTI
+317,113,384,116,383,137,316,134,PLAN
+245,118,310,121,306,136,242,134,###
+443,230,680,208,675,270,438,291,###
+696,212,860,204,857,261,694,268,###
+383,192,398,192,398,208,383,208,###
+216,131,242,455,226,532,188,136,###
+331,550,360,546,361,554,332,557,###
+363,545,416,537,417,544,364,552,###
+419,536,427,535,428,543,420,544,###
+429,534,464,528,467,535,429,542,###
+431,545,455,537,468,546,448,558,###
+588,350,601,349,601,355,588,357,###
+787,326,800,327,800,333,787,336,###
+388,108,428,110,428,121,388,121,###

datasets/icdar2015/test_gts/gt_img_107.txt

@@ -0,0 +1,14 @@
+614,159,646,159,645,173,606,176,@B1
+260,55,475,33,482,85,259,99,MARINA
+516,122,565,122,568,137,518,141,More
+612,111,698,104,702,122,613,128,MARINA
+614,127,698,119,698,153,607,160,LINK
+500,33,705,13,705,63,502,79,SQUARE
+524,146,598,142,601,156,524,159,Shopping
+526,163,597,158,598,178,525,179,&Dining!
+158,325,171,322,171,330,159,332,###
+14,342,34,340,33,347,13,349,###
+32,340,42,340,44,347,34,347,###
+44,337,64,336,66,345,45,347,###
+65,336,82,334,81,344,64,346,###
+1,343,10,343,10,351,1,351,###

datasets/icdar2015/test_gts/gt_img_108.txt

@@ -0,0 +1,16 @@
+655,144,697,139,695,158,652,164,Link
+545,158,572,153,569,176,545,178,B1,
+353,179,367,181,370,203,350,199,###
+374,178,437,170,436,194,375,198,dining
+444,170,515,162,514,187,442,193,options
+252,193,345,178,346,207,248,216,shopping
+520,161,542,161,541,178,521,180,###
+192,200,246,193,247,215,189,219,More
+583,153,650,147,648,165,582,172,Marina
+349,226,433,218,433,258,345,268,This
+442,216,521,212,521,247,443,260,way
+523,205,566,204,568,240,527,244,###
+283,283,497,258,492,292,283,323,Esplanade
+504,250,637,244,637,271,500,292,Station
+226,442,329,429,329,447,227,461,MARINA:
+328,429,404,416,406,436,329,446,SQUARE

datasets/icdar2015/test_gts/gt_img_109.txt

@@ -0,0 +1,16 @@
+578,222,667,241,661,296,577,288,###
+452,205,561,225,558,286,453,274,LOVE
+142,350,227,348,231,381,142,381,###
+154,381,220,384,215,411,154,414,###
+128,414,242,410,248,447,131,458,Organto
+335,354,386,355,386,384,334,383,LOVE
+400,350,444,357,446,384,400,383,###
+664,483,724,476,728,497,667,508,###
+325,411,381,407,378,421,323,424,###
+305,430,377,423,375,437,303,444,###
+288,385,435,379,435,395,286,403,###
+555,237,582,240,583,275,555,274,###
+598,300,621,297,622,311,598,314,###
+858,339,898,339,924,495,871,514,###
+3,384,19,385,22,405,2,403,###
+385,363,399,363,400,382,385,382,###

datasets/icdar2015/test_gts/gt_img_11.txt

@@ -0,0 +1,4 @@
+404,64,465,56,466,71,404,79,BEWARE
+468,54,489,52,490,67,469,69,###
+393,83,502,67,503,83,393,99,MAINTENANCE
+415,99,483,90,484,107,416,116,VEICHLES

datasets/icdar2015/test_gts/gt_img_110.txt

@@ -0,0 +1,10 @@
+1,380,71,371,74,416,4,431,STEP
+74,366,94,360,96,408,75,416,###
+4,432,101,405,105,450,6,476,CHOOSE
+3,477,76,456,80,505,4,527,YOUR
+76,453,174,424,178,471,78,502,TOPPINGS
+750,155,795,151,795,176,750,181,ORO
+794,160,846,156,846,171,794,174,###
+876,201,1016,193,1016,216,876,224,###
+876,223,1022,224,1022,248,876,247,###
+953,250,1021,251,1020,281,952,280,###

datasets/icdar2015/test_gts/gt_img_111.txt

@@ -0,0 +1,9 @@
+157,132,221,151,220,195,158,185,DEFY
+222,150,288,171,286,207,222,194,EMPIRE
+171,448,237,426,240,471,176,500,DEFY
+234,421,303,402,308,441,237,471,ENA E
+361,375,416,361,424,390,365,415,###
+177,272,220,276,218,304,181,306,###
+175,298,224,298,230,323,178,326,###
+1180,15,1262,3,1275,66,1193,78,###
+349,196,400,204,405,238,355,231,###

datasets/icdar2015/test_gts/gt_img_112.txt

@@ -0,0 +1,22 @@
+752,240,802,239,803,260,752,262,Gold
+757,75,802,83,802,98,758,101,###
+743,166,797,163,800,188,742,187,CLASS
+650,82,683,79,683,98,650,101,East
+740,47,782,48,788,71,741,68,VILLAGE
+678,78,722,81,724,99,684,100,Wing
+685,48,738,47,737,64,687,70,GOLDEN
+716,97,765,99,766,114,719,115,Level 3
+726,80,760,81,758,97,725,98,next
+670,164,741,163,741,194,675,195,GOLD
+909,87,964,72,965,101,910,114,HABA
+683,246,735,244,735,267,683,270,grab
+706,317,795,310,797,327,708,333,#GVSuntecCity
+687,337,809,330,810,345,687,352,###
+1103,96,1163,84,1165,110,1105,122,###
+1082,122,1150,111,1152,140,1084,151,###
+1113,148,1157,132,1158,155,1114,172,###
+1180,111,1243,95,1246,120,1183,135,###
+1098,21,1215,1,1225,28,1104,60,LUSH
+1226,3,1258,0,1279,12,1234,27,###
+1235,585,1279,577,1279,598,1235,605,###
+737,245,748,243,749,261,735,264,###

datasets/icdar2015/test_gts/gt_img_113.txt

@@ -0,0 +1,20 @@
+240,533,257,523,257,537,241,546,###
+226,271,285,272,283,287,224,286,###
+413,313,447,301,448,372,417,384,###
+465,18,531,77,525,131,467,85,###
+240,334,300,326,300,372,242,384,SALE
+125,335,173,340,176,365,130,371,SAL
+226,396,292,396,296,435,228,455,###
+217,221,282,223,283,277,221,274,SALE
+1112,70,1174,46,1174,112,1112,127,###
+95,637,142,636,138,657,92,658,sale
+231,451,293,430,293,445,231,466,###
+258,523,284,507,285,521,259,537,###
+285,506,294,502,295,513,286,518,###
+243,548,293,517,293,532,243,562,###
+64,174,90,172,88,190,63,192,###
+90,171,109,169,110,187,90,189,###
+111,170,128,169,130,185,112,187,###
+131,166,140,165,140,185,130,185,###
+562,128,600,155,593,188,555,162,###
+623,231,678,230,678,245,623,246,###

datasets/icdar2015/test_gts/gt_img_114.txt

@@ -0,0 +1,14 @@
+412,660,480,671,476,718,415,717,H&M
+543,247,701,236,697,272,545,281,###
+1086,497,1226,504,1231,558,1088,554,FINAL
+1237,510,1272,515,1273,561,1240,562,###
+825,550,1058,561,1066,647,832,636,SALE
+948,502,1081,510,1082,556,952,548,FINAL
+1072,563,1267,570,1276,665,1080,660,SAL
+820,501,942,503,942,548,820,542,FINAL
+31,345,64,344,62,364,28,365,###
+65,344,126,345,126,364,65,363,FARN
+294,342,318,342,318,364,294,364,###
+301,220,341,221,340,238,300,237,###
+42,208,87,210,86,227,41,226,###
+220,217,247,217,247,231,220,231,###

datasets/icdar2015/test_gts/gt_img_115.txt

@@ -0,0 +1,6 @@
+102,330,152,327,163,345,106,356,SWEATS
+104,358,183,345,187,364,106,382,COMFORTABLE
+182,336,250,330,247,350,185,363,YERSATIUITT
+107,385,161,372,160,388,111,401,DEFINED
+793,60,862,58,864,74,795,75,HARA
+863,54,917,57,918,78,864,75,###

datasets/icdar2015/test_gts/gt_img_116.txt

@@ -0,0 +1,7 @@
+391,196,442,176,450,215,393,227,FOSSIL
+814,212,862,207,865,250,818,256,UNI
+817,251,870,253,867,303,816,308,QLO
+397,542,447,581,444,613,394,574,###
+293,235,315,232,314,254,292,257,###
+293,450,319,438,319,456,293,468,###
+4,283,17,282,14,320,2,322,###

datasets/icdar2015/test_gts/gt_img_117.txt

@@ -0,0 +1,7 @@
+30,71,120,108,116,154,32,116,###
+751,68,873,0,882,55,762,118,SELANGOR
+971,304,1026,306,1026,325,971,323,###
+962,332,1037,341,1038,357,963,348,###
+292,253,324,250,324,265,292,268,###
+767,32,823,3,827,26,770,55,###
+292,388,324,396,324,419,292,411,###

datasets/icdar2015/test_gts/gt_img_118.txt

@@ -0,0 +1,6 @@
+922,220,1021,208,1016,235,922,245,TISSOT
+1141,196,1215,207,1212,235,1145,230,TISSOT
+723,242,852,245,852,266,723,263,###
+53,258,108,251,108,271,53,278,###
+122,247,156,242,155,261,121,266,###
+170,241,218,228,217,245,168,257,###

datasets/icdar2015/test_gts/gt_img_119.txt

@@ -0,0 +1,6 @@
+10,46,178,81,183,132,15,116,Robert
+177,80,301,114,300,152,184,133,Timms
+305,110,350,117,348,134,304,126,Robert
+306,126,346,131,346,150,306,145,Timms
+90,168,154,175,152,200,87,193,###
+136,381,193,377,192,404,135,407,###

datasets/icdar2015/test_gts/gt_img_12.txt

@@ -0,0 +1,16 @@
+122,428,163,425,160,442,119,445,###
+1146,105,1220,92,1240,135,1166,148,###
+95,223,134,220,132,237,93,240,those
+42,248,77,240,80,277,45,285,###
+81,243,169,238,160,277,80,273,need!
+69,180,142,173,139,216,66,223,Care
+66,225,97,220,96,243,65,248,for
+80,409,107,410,108,425,81,424,Look
+332,146,407,140,408,157,333,163,###
+67,435,84,435,87,445,69,446,###
+88,429,117,429,117,445,87,445,Care
+26,313,59,315,57,330,24,329,###
+77,330,122,330,122,348,77,348,###
+146,305,177,306,180,325,150,324,###
+148,360,178,361,179,380,150,379,###
+29,359,59,360,63,378,33,377,###

datasets/icdar2015/test_gts/gt_img_120.txt

@@ -0,0 +1,16 @@
+1094,387,1144,391,1145,410,1095,406,###
+1091,350,1143,352,1143,372,1091,370,###
+560,307,632,305,630,326,560,323,###
+771,244,877,235,876,255,773,261,###
+73,232,125,235,125,243,73,240,###
+637,30,705,1,752,15,650,77,###
+1050,352,1086,352,1087,371,1051,371,###
+43,208,156,214,153,238,47,231,SINCERE
+1054,387,1094,390,1095,406,1055,404,###
+1144,388,1190,394,1191,414,1145,408,###
+398,301,433,301,433,315,398,315,###
+518,226,550,228,550,251,518,248,###
+1155,349,1202,345,1207,383,1161,386,###
+1212,311,1256,311,1256,357,1212,357,###
+1189,100,1215,100,1215,130,1189,130,###
+0,663,144,630,148,680,4,713,###

datasets/icdar2015/test_gts/gt_img_121.txt

@@ -0,0 +1,21 @@
+916,128,1012,57,1018,114,923,184,ESPRIT
+344,266,408,266,407,283,344,286,###
+597,223,646,221,651,247,597,248,Enjoy
+590,248,646,244,643,280,594,283,great
+647,236,714,250,711,276,648,280,deals
+598,177,717,172,714,217,597,213,privileges
+596,283,675,275,676,310,598,310,Suntec
+590,144,722,134,718,178,591,181,exciting
+248,122,292,157,284,190,243,165,SINCERE
+674,268,715,278,720,311,677,305,City
+603,405,718,402,721,427,602,427,###
+601,365,718,365,718,376,600,382,###
+594,376,723,377,723,394,596,392,###
+20,160,52,165,50,202,17,196,###
+761,313,777,314,777,328,761,327,###
+764,432,783,432,783,447,764,447,###
+242,295,284,296,284,305,242,304,###
+604,118,627,116,627,131,604,133,###
+626,116,652,115,652,127,626,128,that
+653,113,685,115,683,130,651,127,gives
+684,115,706,116,705,132,683,131,you

datasets/icdar2015/test_gts/gt_img_122.txt

@@ -0,0 +1,22 @@
+976,171,1004,176,1007,217,973,219,###
+519,221,547,221,547,234,517,234,###
+1008,112,1086,96,1090,132,1008,144,SALE
+1115,26,1163,10,1173,23,1112,41,FURTHER
+1106,44,1179,23,1181,44,1107,62,REDUCTIONS
+665,159,731,163,733,185,667,181,GIORDAN
+911,191,971,179,978,242,913,238,###
+763,161,823,151,830,221,766,228,GAP
+210,268,253,267,253,307,212,309,###
+963,83,1013,71,1017,89,963,98,###
+1162,79,1246,65,1247,106,1172,114,###
+970,66,1004,56,1008,72,969,82,###
+975,220,1010,218,1005,236,978,237,OFF
+1056,165,1132,152,1146,231,1059,231,###
+1134,155,1185,152,1193,202,1147,206,###
+1140,206,1196,206,1175,228,1147,227,OFF
+1152,335,1203,339,1203,354,1152,351,###
+1205,340,1232,345,1231,357,1206,356,COM
+1234,343,1262,343,1258,362,1234,359,###
+783,462,855,492,851,548,779,519,###
+1039,150,1225,139,1225,149,1040,160,###
+911,163,975,163,975,171,911,171,###

datasets/icdar2015/test_gts/gt_img_123.txt

@@ -0,0 +1,4 @@
+216,128,360,173,358,207,217,173,GIORDANO
+1197,0,1277,0,1279,151,1209,146,###
+522,241,540,243,540,258,522,256,###
+658,155,671,156,670,191,657,190,###

datasets/icdar2015/test_gts/gt_img_124.txt

@@ -0,0 +1,7 @@
+714,70,759,43,774,104,718,123,###
+546,162,579,140,581,173,548,188,###
+193,150,231,165,227,190,192,180,###
+84,173,160,190,156,213,85,203,###
+728,277,757,277,756,302,727,299,###
+240,220,306,222,308,238,242,236,###
+252,242,310,243,308,257,251,256,###

datasets/icdar2015/test_gts/gt_img_125.txt

@@ -0,0 +1,34 @@
+306,307,323,308,322,323,305,323,###
+91,309,121,301,123,338,93,343,sale
+304,181,339,187,340,201,303,201,cafes
+224,205,250,203,251,218,223,218,and
+254,203,335,203,337,216,254,219,Restaurants
+255,186,300,190,300,202,252,202,shops
+232,227,267,227,267,242,226,242,food
+215,189,251,191,250,202,211,202,more
+271,223,330,225,331,239,269,244,republic
+219,244,260,247,260,257,215,259,giant
+264,239,342,243,342,260,262,264,hyperfresh
+220,262,280,270,280,283,211,285,fountain
+283,268,298,264,299,279,281,281,###
+298,265,335,264,335,280,299,282,###
+210,314,260,313,261,326,206,331,suntec
+226,287,277,286,278,300,227,302,Money
+278,283,322,281,322,300,276,300,Chang
+265,309,304,309,306,323,262,327,office
+124,304,148,302,149,332,123,335,SALE
+150,299,173,303,174,327,147,331,SALE
+766,99,837,55,840,99,777,129,COTTON
+844,48,881,32,884,81,844,97,###
+812,245,825,244,828,269,814,270,###
+1015,265,1032,267,1033,302,1018,298,###
+1052,263,1081,269,1083,308,1056,308,###
+991,143,1029,153,1030,183,989,190,new
+994,173,1033,183,1034,213,989,224,NOW
+52,0,163,61,158,91,2,16,###
+85,190,137,187,140,206,89,208,###
+90,209,127,210,127,225,90,224,###
+89,237,121,235,120,249,87,250,###
+84,249,129,247,132,267,87,268,###
+89,272,141,272,141,291,89,291,###
+90,297,130,292,131,305,90,310,###

datasets/icdar2015/test_gts/gt_img_126.txt

@@ -0,0 +1,15 @@
+270,169,319,199,320,247,274,224,dining
+230,154,274,168,271,219,234,203,and
+264,102,326,143,324,177,262,148,continues
+1,48,77,81,61,153,0,139,###
+80,71,236,156,229,211,78,155,shopping
+172,31,272,101,259,147,172,98,remaking
+104,0,182,1,160,82,124,75,city
+4,224,95,236,96,298,3,316,you
+94,221,152,228,145,305,106,296,for
+150,250,210,256,210,296,150,316,your
+210,256,290,266,288,304,212,319,support
+53,0,101,1,103,29,72,17,###
+1113,264,1146,260,1147,300,1113,305,###
+927,192,967,191,970,255,928,257,###
+960,286,1010,284,1014,340,964,341,###

datasets/icdar2015/test_gts/gt_img_127.txt

@@ -0,0 +1,9 @@
+1159,0,1268,0,1266,9,1161,19,SPECIAL
+430,22,574,104,575,153,447,96,robinsons
+452,142,514,140,516,168,450,172,SK-II
+606,156,619,157,618,179,609,178,###
+961,150,1012,145,1012,155,961,160,###
+981,155,1012,155,1012,164,981,164,###
+575,136,618,133,620,146,576,150,###
+186,154,205,155,206,178,187,177,###
+116,166,167,161,170,183,118,188,###

datasets/icdar2015/test_gts/gt_img_128.txt

@@ -0,0 +1,3 @@
+1135,67,1246,56,1246,74,1138,84,SINCLARE
+1023,2,1157,0,1158,29,1008,68,###
+708,120,735,122,734,134,709,132,###

datasets/icdar2015/test_gts/gt_img_129.txt

@@ -0,0 +1,10 @@
+920,554,1000,550,1003,581,925,590,raffles
+1002,547,1053,548,1055,578,1001,583,City
+1148,173,1162,172,1164,197,1146,201,###
+1192,161,1219,153,1223,165,1194,172,###
+1236,183,1260,179,1263,193,1236,197,###
+1242,131,1277,119,1278,136,1246,143,###
+1192,178,1242,166,1243,177,1196,190,###
+1188,145,1241,131,1243,145,1191,159,###
+1205,191,1222,186,1222,200,1203,205,###
+0,646,84,636,87,694,2,702,###

datasets/icdar2015/test_gts/gt_img_13.txt

@@ -0,0 +1,20 @@
+3,159,14,159,15,172,1,175,###
+290,57,333,43,336,51,292,67,###
+805,233,1052,176,1062,211,815,268,CHEVRON
+1114,93,1153,82,1158,96,1119,107,###
+35,279,69,267,68,289,35,300,###
+367,204,409,192,410,205,368,217,Premier
+927,145,1088,103,1100,120,923,164,###
+328,192,375,177,378,197,331,211,HSBC
+284,29,368,1,375,9,289,38,###
+286,44,364,19,366,26,286,53,###
+292,71,313,65,319,72,293,81,###
+14,176,38,164,44,178,16,192,###
+18,184,58,170,69,188,20,208,###
+13,155,56,140,60,156,15,169,###
+293,95,373,65,379,95,296,120,###
+375,260,506,230,507,250,379,283,###
+412,279,475,259,475,280,419,291,###
+1060,172,1194,134,1212,181,1072,217,HOU
+1117,106,1163,94,1169,106,1112,120,###
+623,26,667,24,669,38,625,39,###

datasets/icdar2015/test_gts/gt_img_130.txt

@@ -0,0 +1,10 @@
+180,51,287,130,280,181,172,102,ROBINSONS
+187,183,256,188,254,205,185,200,###
+8,186,93,193,91,222,6,215,###
+405,102,446,114,443,130,402,117,THT
+684,238,732,223,732,241,684,256,###
+820,167,864,148,862,171,817,190,###
+911,274,932,268,931,286,910,292,###
+337,161,364,180,362,207,335,188,###
+385,217,403,231,398,257,381,244,###
+384,331,403,332,401,347,382,346,###

datasets/icdar2015/test_gts/gt_img_131.txt

@@ -0,0 +1,8 @@
+625,213,716,212,715,237,626,243,MARKET
+717,217,780,223,781,244,716,246,PLACE
+936,358,991,353,992,374,936,376,cierge
+204,371,278,365,283,390,205,390,BRITISH
+742,343,789,330,793,354,746,358,###
+787,385,802,383,802,396,789,394,###
+785,396,803,394,806,409,788,408,###
+37,335,73,341,72,355,36,350,###

datasets/icdar2015/test_gts/gt_img_132.txt

@@ -0,0 +1,11 @@
+68,305,106,304,102,328,68,326,###
+374,195,481,209,482,235,374,222,SACOOR
+487,217,538,224,542,242,487,236,###
+634,236,652,237,653,252,634,248,###
+656,236,688,234,686,257,656,252,###
+688,242,731,248,732,263,688,258,###
+793,278,877,276,882,296,792,299,BRITISH
+593,231,631,233,631,247,593,245,###
+62,257,126,261,126,281,62,277,###
+1196,0,1278,1,1278,22,1193,25,###
+565,113,682,123,688,138,566,132,###

datasets/icdar2015/test_gts/gt_img_133.txt

@@ -0,0 +1,20 @@
+931,191,984,182,983,205,932,208,###
+978,265,1033,264,1034,281,979,281,SHOPS
+220,102,341,157,331,181,214,130,jacobs
+471,418,506,421,505,435,470,433,SALE
+452,290,608,288,609,321,458,323,BRITISH
+101,44,208,95,202,117,90,67,MARC
+611,283,738,281,738,323,609,320,INDIA
+15,4,79,34,68,60,4,33,###
+794,427,894,428,896,471,788,465,SALE
+932,268,976,265,976,281,934,281,###
+927,284,1044,278,1043,301,936,301,RESTAURANTS
+986,185,1042,187,1042,204,987,203,###
+1042,184,1062,185,1064,204,1045,203,###
+987,157,1063,157,1063,177,987,171,###
+924,166,982,155,982,170,934,175,###
+925,150,992,137,993,156,931,161,RAFFLES
+995,141,1031,136,1029,153,998,153,###
+1028,140,1076,143,1074,160,1036,155,###
+1167,442,1279,428,1277,493,1156,484,###
+284,280,315,281,314,298,283,297,###

datasets/icdar2015/test_gts/gt_img_134.txt

@@ -0,0 +1,3 @@
+514,390,614,396,610,432,518,426,SALE
+484,56,932,2,930,92,464,138,BRITISH
+942,0,1279,4,1276,48,944,78,###

datasets/icdar2015/test_gts/gt_img_135.txt

@@ -0,0 +1,3 @@
+371,413,451,412,450,432,371,428,dont
+364,432,455,435,454,456,365,459,panic
+890,194,1012,217,1008,240,886,216,###

datasets/icdar2015/test_gts/gt_img_136.txt

@@ -0,0 +1,9 @@
+50,224,112,224,114,250,46,238,place
+226,34,284,28,284,124,242,122,bright
+135,320,175,315,176,328,136,333,###
+343,345,391,338,412,442,364,448,###
+348,455,400,473,362,596,311,578,###
+306,464,341,475,315,547,281,536,###
+0,224,51,222,48,239,0,240,RKET
+213,0,263,0,264,16,229,21,###
+293,0,336,0,324,66,286,59,WALK

datasets/icdar2015/test_gts/gt_img_137.txt

@@ -0,0 +1,11 @@
+494,198,580,158,592,186,508,218,JACOBS
+6,388,20,382,41,415,6,439,###
+306,278,378,240,390,270,314,300,MARC
+382,242,418,224,434,252,400,264,###
+424,226,486,194,498,224,440,246,MARC
+59,385,169,334,180,357,71,405,brothers
+866,76,934,54,944,72,886,87,DIA
+732,108,758,102,765,120,734,128,###
+800,245,872,236,875,260,808,271,###
+804,272,874,266,876,285,808,290,###
+816,314,868,317,868,336,822,335,###

datasets/icdar2015/test_gts/gt_img_138.txt

@@ -0,0 +1,12 @@
+40,134,254,116,258,166,52,186,desigual
+433,156,447,155,452,192,438,193,###
+312,170,386,163,389,183,320,190,desigual
+726,164,845,149,864,240,749,253,###
+995,130,1058,119,1065,143,1003,150,alluti
+452,86,646,70,656,114,466,128,###
+238,289,285,285,291,316,242,322,###
+477,156,544,148,548,169,484,175,###
+480,176,562,162,566,184,486,190,###
+184,194,261,186,262,205,191,214,Further
+232,208,287,201,291,218,235,225,tions
+821,88,898,106,895,135,817,117,###

datasets/icdar2015/test_gts/gt_img_139.txt

@@ -0,0 +1,2 @@
+331,115,449,113,457,156,329,156,PRECIOUS
+451,114,560,107,560,157,459,157,THOTS

datasets/icdar2015/test_gts/gt_img_14.txt

@@ -0,0 +1,6 @@
+268,82,335,93,332,164,267,164,the
+344,94,433,112,427,159,336,163,Future
+208,191,374,184,371,213,208,241,Communications
+370,176,420,176,416,204,373,213,###
+1,57,261,76,261,187,0,190,venting
+1,208,203,200,203,241,3,294,ntelligence.

datasets/icdar2015/test_gts/gt_img_140.txt

@@ -0,0 +1,16 @@
+783,397,809,399,809,411,788,414,###
+821,452,836,450,835,464,823,464,###
+578,366,603,366,602,389,577,389,###
+618,434,659,436,656,451,619,453,###
+616,417,660,416,659,436,615,433,SUSHI
+228,320,291,324,286,336,229,329,###
+530,363,578,365,576,390,531,387,sushi
+785,355,807,356,805,396,785,394,the
+1180,374,1250,373,1252,398,1181,398,kimmic
+900,230,1004,162,1010,228,900,285,precious
+994,108,1145,103,1154,144,1004,230,###
+706,314,746,291,738,336,710,346,###
+729,298,811,281,813,324,737,348,place
+510,603,590,612,588,637,508,628,###
+620,364,662,366,659,419,617,417,###
+624,363,658,365,657,417,622,415,###

datasets/icdar2015/test_gts/gt_img_141.txt

@@ -0,0 +1,22 @@
+766,114,1125,2,1170,47,772,188,###
+882,685,904,691,905,719,879,717,###
+1037,460,1117,465,1126,530,1056,530,###
+1052,531,1091,528,1092,546,1057,546,###
+1092,525,1147,526,1146,550,1097,547,###
+707,245,755,235,760,253,710,267,Noodles
+674,221,748,198,750,225,677,246,PONTIAN
+668,259,708,250,710,265,672,276,Wanton
+961,511,1024,506,1025,526,965,527,###
+922,381,1146,357,1156,440,931,451,###
+666,150,750,112,752,165,669,202,###
+686,191,738,173,742,201,690,219,###
+958,462,1021,465,1021,486,958,482,###
+968,492,1017,494,1021,512,972,510,###
+783,220,869,191,873,206,787,236,###
+782,234,831,225,831,238,782,247,###
+932,171,986,155,991,171,937,187,###
+931,190,1060,152,1063,172,935,210,###
+776,216,875,188,885,222,779,253,###
+774,190,1095,80,1107,116,776,215,###
+929,171,1060,131,1069,174,927,209,###
+845,85,1051,0,1066,15,850,93,###

datasets/icdar2015/test_gts/gt_img_142.txt

@@ -0,0 +1,48 @@
+812,272,841,270,842,282,815,282,###
+751,270,782,269,781,283,751,281,###
+554,236,616,234,617,285,551,283,###
+617,228,703,234,712,262,616,259,OPE
+616,258,706,259,705,286,617,284,EARLY
+506,236,529,238,529,287,506,288,###
+576,321,648,326,648,342,575,342,OPENS
+380,239,478,236,478,284,374,284,toast
+559,334,584,338,582,365,565,365,###
+588,339,618,339,612,363,588,362,###
+618,340,657,342,657,365,610,364,###
+782,245,809,242,810,259,785,257,###
+592,362,636,362,635,377,593,377,DAILY
+754,244,784,242,784,256,754,255,###
+832,244,878,242,878,258,834,257,###
+881,246,900,242,902,258,878,258,###
+807,246,833,243,833,258,811,257,###
+902,245,917,246,918,260,904,258,###
+777,258,808,257,808,270,782,269,###
+808,258,832,256,832,270,808,269,###
+748,256,779,256,779,270,752,269,###
+831,259,854,258,856,271,836,270,###
+854,259,880,258,880,272,856,272,###
+879,260,903,258,906,273,878,272,and
+781,270,812,271,813,284,782,282,###
+748,284,780,283,779,296,752,294,###
+778,284,804,283,801,295,780,295,your
+802,284,830,282,829,296,804,296,###
+830,282,880,281,880,295,832,294,important
+880,282,906,278,904,295,883,295,###
+904,283,918,279,918,295,907,294,###
+750,296,769,292,771,308,750,307,###
+772,296,790,294,789,307,771,306,###
+790,296,812,296,813,309,792,307,###
+812,297,826,297,829,308,814,308,###
+1230,249,1252,248,1254,258,1232,260,###
+1210,250,1231,249,1232,260,1211,261,###
+1199,250,1208,249,1208,260,1199,261,###
+1211,260,1234,258,1234,270,1211,272,###
+1233,261,1256,260,1256,271,1233,272,###
+1205,274,1234,270,1236,281,1207,286,###
+1233,275,1255,272,1255,281,1234,285,###
+1215,285,1232,285,1233,293,1215,294,###
+1233,284,1255,283,1255,292,1234,293,###
+1208,297,1225,294,1225,305,1208,307,###
+1226,295,1257,295,1258,306,1228,306,###
+1198,285,1215,285,1215,295,1198,295,###
+1197,261,1209,261,1210,270,1198,270,###