ctd数据格式转普通坐标点

ctd_label2pt.py

import numpy as np

import cv2
import random

import os


def get_absolute_path(p):
    if p.startswith('~'):
        p = os.path.expanduser(p)
    return os.path.abspath(p)

def read_lines(p):
    """return the text in a file in lines as a list """
    p = get_absolute_path(p)
    f = open(p,'r')
    return f.readlines()

def replace_all(s, old, new, reg = False):
    if reg:
        import re
        targets = re.findall(old, s)
        for t in targets:
            s = s.replace(t, new)
    else:
        s = s.replace(old, new)
    return s

def remove_all(s, sub):
    return replace_all(s, sub, '')

def split(s, splitter, reg = False):
    if not reg:
        return s.split(splitter)
    import re
    return re.split(splitter, s)

def get_bboxes(img, gt_path, save_path):
    h, w = img.shape[0:2]
    lines = read_lines(gt_path)
    bboxes = []
    tags = []
    for line in lines:
        line = line.strip()
        line = remove_all(line, 'xefxbbxbf')
        gt = split(line, ',')

        x1 = np.int(gt[0])
        y1 = np.int(gt[1])

        bbox = [np.int(gt[i]) for i in range(4, 32)]
        bbox = np.asarray(bbox) + ([x1 * 1.0, y1 * 1.0] * 14)
        with open(save_path,'a') as fw:
            for cnt,val in enumerate(bbox):
                val = int(val)
                if cnt != len(bbox)-1:
                    fw.write(str(val))
                    fw.write(",")
                else:
                    fw.write(str(val))
                    fw.write('
')


# def get_bboxes(img, gt_path, save_path):
#     h, w = img.shape[0:2]
#     lines = util.io.read_lines(gt_path)
#     bboxes = []
#     tags = []
#     for line in lines:
#         line = line.strip()
#         line = util.str.remove_all(line, 'xefxbbxbf')
#         gt = util.str.split(line, ',')
#
#         x1 = np.int(gt[0])
#         y1 = np.int(gt[1])
#
#         bbox = [np.int(gt[i]) for i in range(4, 32)]
#         bbox = np.asarray(bbox) + ([x1 * 1.0, y1 * 1.0] * 14)
#         with open(save_path,'a') as fw:
#             for cnt,val in enumerate(bbox):
#                 val = int(val)
#                 if cnt != len(bbox)-1:
#                     fw.write(str(val))
#                     fw.write(",")
#                 else:
#                     fw.write(str(val))
#                     fw.write('
')


path_text_image = "/media/data_1/2019_project_test/PSENet/data/CTW1500/train/text_image/"
path_text_label_curve = "/media/data_1/2019_project_test/PSENet/data/CTW1500/train/text_label_curve/"
path_save_txt = os.path.dirname(os.path.dirname(path_text_image)) + '/ctd2general/'
if not os.path.exists(path_save_txt):
    os.makedirs(path_save_txt)

list_img = os.listdir(path_text_image)
cnt = 0
for img_name in list_img:
    cnt += 1
    print("cnt=%d,img=%s"%(cnt,img_name))
    img_path = path_text_image + img_name
    img = cv2.imread(img_path)
    txt_path = path_text_label_curve + img_name.replace('.jpg','.txt')
    txt_save = path_save_txt + img_name.replace('.jpg','.txt')
    get_bboxes(img, txt_path, txt_save)