""" Walk-Assistant : Recognizing sidewalk for the visually impaired Copyright (C) 2018 Yoongi Kim (devlifecode@outlook.com) This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program. If not, see <http://www.gnu.org/licenses/>. """ print(""" Walk-Assistant Copyright (C) 2018 Yoongi Kim This program comes with ABSOLUTELY NO WARRANTY. This is free software, and you are welcome to redistribute it under certain conditions. """) import cv2 from data_loader import DataLoader import glob import numpy as np import random import os from model import MyModel PATH = "data/frames" ANNOTATION_FILE = "data/label.txt" KERNEL = 80 SHAPE = (9, 16) HEIGHT = 720 WIDTH = 1280 labels = [] grid = np.zeros(SHAPE, dtype=np.int) l_down = False r_down = False def click(event, x, y, flags, param): global clicks, l_down, r_down if event == cv2.EVENT_LBUTTONDOWN or (event == cv2.EVENT_MOUSEMOVE and l_down): l_down = True x = max(0, min(x, WIDTH-1)) y = max(0, min(y, HEIGHT-1)) r = y // KERNEL c = x // KERNEL grid[r][c] = 1 elif event == cv2.EVENT_LBUTTONUP: l_down = False if event == cv2.EVENT_RBUTTONDOWN or (event == cv2.EVENT_MOUSEMOVE and r_down): r_down = True r = y // KERNEL c = x // KERNEL grid[r][c] = 0 elif event == cv2.EVENT_RBUTTONUP: r_down = False def visualize(img): global grid grid = np.round(grid) h, w, c = img.shape box = np.zeros((h, w), dtype=np.uint8) for i in range(len(grid)): for j in range(len(grid[i])): box[i*KERNEL : (i+1)*KERNEL, j*KERNEL : (j+1)*KERNEL] = grid[i][j] * 255 box = cv2.cvtColor(box, cv2.COLOR_GRAY2BGR) box[:,:,0]=0 box[:,:,2]=0 visual = cv2.add(img, box) for i in range(0, HEIGHT, KERNEL): # 가로선 cv2.line(visual, (0, i), (WIDTH, i), color=(255,255,255)) for i in range(0, WIDTH, KERNEL): # 세로선 cv2.line(visual, (i, 0), (i, HEIGHT), color=(255,255,255)) return visual def read_label(): files = [] if not os.path.exists(ANNOTATION_FILE): return [], [] with open(ANNOTATION_FILE, 'r') as f: lines = f.readlines() for line in lines: line = str(line).replace('\n', '') file, encode = line.split(',') files.append(file) return files, lines def write_label(file_name, grid): grid = np.round(grid).astype(np.uint8) flat = np.array(grid).flatten() encode = '' label = '' for b in flat: encode += str(b) label = '%s,%s\n' % (file_name, encode) print(label) labels.append(label) with open(ANNOTATION_FILE, 'w') as f: f.writelines(labels) print('Saved label file') def predict(img, my_model): img = np.array([img])/255.0 res = my_model.model.predict(img)[0] res = np.squeeze(res[:,:,1]) # (9, 16) return res if __name__ == '__main__': model = MyModel(True, HEIGHT, WIDTH, KERNEL, KERNEL, 1e-6, 'main') files = glob.glob('{}/*.jpg'.format(PATH)) prev_files, prev_lines = read_label() labels = prev_lines print('Previous annotations: {}'.format(len(prev_files))) print(prev_files) random.shuffle(files) for index, file in enumerate(files): file_name = str(file).replace('\\', '/').split('/')[-1] if file_name in prev_files: continue img = cv2.imread(file) grid = predict(img, model) cv2.namedWindow('img') cv2.setMouseCallback('img', click) while True: cv2.imshow('img', visualize(img)) k = cv2.waitKey(2) if k == -1: # No Key pass elif k == 27: # ESC print('Exit') exit(0) elif k == 32 or k == 13: # ESC, Space, Enter write_label(file_name, grid) break else: print('Key pressed: ', k) print('Done')