predict_seuil_expe_maxwell.py 8.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223
  1. from sklearn.externals import joblib
  2. import numpy as np
  3. from ipfml import image_processing
  4. from PIL import Image
  5. import sys, os, getopt
  6. import subprocess
  7. import time
  8. current_dirpath = os.getcwd()
  9. config_filename = "config"
  10. scenes_path = './fichiersSVD_light'
  11. min_max_filename = '_min_max_values'
  12. threshold_expe_filename = 'seuilExpe'
  13. tmp_filename = '/tmp/__model__img_to_predict.png'
  14. maxwell_scenes = ['Appart1opt02', 'Cuisine01', 'SdbCentre', 'SdbDroite']
  15. threshold_map_folder = "threshold_map"
  16. threshold_map_file_prefix = "treshold_map_"
  17. zones = np.arange(16)
  18. def main():
  19. if len(sys.argv) <= 1:
  20. print('Run with default parameters...')
  21. print('python predict_seuil_expe_maxwell.py --interval "0,20" --model path/to/xxxx.joblib --mode svdn --metric lab --limit_detection xx')
  22. sys.exit(2)
  23. try:
  24. opts, args = getopt.getopt(sys.argv[1:], "ht:m:o:l", ["help=", "interval=", "model=", "mode=", "metric=", "limit_detection="])
  25. except getopt.GetoptError:
  26. # print help information and exit:
  27. print('python predict_seuil_expe_maxwell.py --interval "xx,xx" --model path/to/xxxx.joblib --mode svdn --metric lab --limit_detection xx')
  28. sys.exit(2)
  29. for o, a in opts:
  30. if o == "-h":
  31. print('python predict_seuil_expe_maxwell.py --interval "xx,xx" --model path/to/xxxx.joblib --mode svdn --metric lab --limit_detection xx')
  32. sys.exit()
  33. elif o in ("-t", "--interval"):
  34. p_interval = a
  35. elif o in ("-m", "--model"):
  36. p_model_file = a
  37. elif o in ("-o", "--mode"):
  38. p_mode = a
  39. if p_mode != 'svdn' and p_mode != 'svdne' and p_mode != 'svd':
  40. assert False, "Mode not recognized"
  41. elif o in ("-m", "--metric"):
  42. p_metric = a
  43. elif o in ("-l", "--limit_detection"):
  44. p_limit = int(a)
  45. else:
  46. assert False, "unhandled option"
  47. scenes = os.listdir(scenes_path)
  48. if min_max_filename in scenes:
  49. scenes.remove(min_max_filename)
  50. # go ahead each scenes
  51. for id_scene, folder_scene in enumerate(scenes):
  52. # only take in consideration maxwell scenes
  53. if folder_scene in maxwell_scenes:
  54. print(folder_scene)
  55. scene_path = os.path.join(scenes_path, folder_scene)
  56. config_path = os.path.join(scene_path, config_filename)
  57. with open(config_path, "r") as config_file:
  58. last_image_name = config_file.readline().strip()
  59. prefix_image_name = config_file.readline().strip()
  60. start_index_image = config_file.readline().strip()
  61. end_index_image = config_file.readline().strip()
  62. step_counter = int(config_file.readline().strip())
  63. threshold_expes = []
  64. threshold_expes_detected = []
  65. threshold_expes_counter = []
  66. threshold_expes_found = []
  67. # get zones list info
  68. for index in zones:
  69. index_str = str(index)
  70. if len(index_str) < 2:
  71. index_str = "0" + index_str
  72. zone_folder = "zone"+index_str
  73. threshold_path_file = os.path.join(os.path.join(scene_path, zone_folder), threshold_expe_filename)
  74. with open(threshold_path_file) as f:
  75. threshold = int(f.readline())
  76. threshold_expes.append(threshold)
  77. # Initialize default data to get detected model threshold found
  78. threshold_expes_detected.append(False)
  79. threshold_expes_counter.append(0)
  80. threshold_expes_found.append(int(end_index_image)) # by default use max
  81. current_counter_index = int(start_index_image)
  82. end_counter_index = int(end_index_image)
  83. print(current_counter_index)
  84. check_all_done = False
  85. while(current_counter_index <= end_counter_index and not check_all_done):
  86. current_counter_index_str = str(current_counter_index)
  87. while len(start_index_image) > len(current_counter_index_str):
  88. current_counter_index_str = "0" + current_counter_index_str
  89. img_path = os.path.join(scene_path, prefix_image_name + current_counter_index_str + ".png")
  90. current_img = Image.open(img_path)
  91. img_blocks = image_processing.divide_in_blocks(current_img, (200, 200))
  92. check_all_done = all(d == True for d in threshold_expes_detected)
  93. for id_block, block in enumerate(img_blocks):
  94. # check only if necessary for this scene (not already detected)
  95. if not threshold_expes_detected[id_block]:
  96. tmp_file_path = tmp_filename.replace('__model__', p_model_file.split('/')[-1].replace('.joblib', '_'))
  97. block.save(tmp_file_path)
  98. python_cmd = "python metrics_predictions/predict_noisy_image_svd_" + p_metric + ".py --image " + tmp_file_path + \
  99. " --interval '" + p_interval + \
  100. "' --model " + p_model_file + \
  101. " --mode " + p_mode
  102. ## call command ##
  103. p = subprocess.Popen(python_cmd, stdout=subprocess.PIPE, shell=True)
  104. (output, err) = p.communicate()
  105. ## Wait for result ##
  106. p_status = p.wait()
  107. prediction = int(output)
  108. if prediction == 0:
  109. threshold_expes_counter[id_block] = threshold_expes_counter[id_block] + 1
  110. else:
  111. threshold_expes_counter[id_block] = 0
  112. if threshold_expes_counter[id_block] == p_limit:
  113. threshold_expes_detected[id_block] = True
  114. threshold_expes_found[id_block] = current_counter_index
  115. print(str(id_block) + " : " + str(current_counter_index) + "/" + str(threshold_expes[id_block]) + " => " + str(prediction))
  116. current_counter_index += step_counter
  117. print("------------------------")
  118. print("Scene " + str(id_scene + 1) + "/" + str(len(scenes)))
  119. print("------------------------")
  120. # end of scene => display of results
  121. # construct path using model name for saving threshold map folder
  122. model_treshold_path = os.path.join(threshold_map_folder, p_model_file.split('/')[-1].replace('.joblib', '_'))
  123. # create threshold model path if necessary
  124. if not os.path.exists(model_treshold_path):
  125. os.makedirs(model_treshold_path)
  126. abs_dist = []
  127. map_filename = os.path.join(model_treshold_path, threshold_map_file_prefix + folder_scene)
  128. f_map = open(map_filename, 'w')
  129. line_information = ""
  130. # default header
  131. f_map.write('| | | | |\n')
  132. f_map.write('---|----|----|---\n')
  133. for id, threshold in enumerate(threshold_expes_found):
  134. line_information += str(threshold) + " / " + str(threshold_expes[id]) + " | "
  135. abs_dist.append(abs(threshold - threshold_expes[id]))
  136. if (id + 1) % 4 == 0:
  137. f_map.write(line_information + '\n')
  138. line_information = ""
  139. f_map.write(line_information + '\n')
  140. min_abs_dist = min(abs_dist)
  141. max_abs_dist = max(abs_dist)
  142. avg_abs_dist = sum(abs_dist) / len(abs_dist)
  143. f_map.write('\nScene information : ')
  144. f_map.write('\n- BEGIN : ' + str(start_index_image))
  145. f_map.write('\n- END : ' + str(end_index_image))
  146. f_map.write('\n\nDistances information : ')
  147. f_map.write('\n- MIN : ' + str(min_abs_dist))
  148. f_map.write('\n- MAX : ' + str(max_abs_dist))
  149. f_map.write('\n- AVG : ' + str(avg_abs_dist))
  150. f_map.write('\n\nOther information : ')
  151. f_map.write('\n- Detection limit : ' + str(p_limit))
  152. # by default print last line
  153. f_map.close()
  154. print("Scene " + str(id_scene + 1) + "/" + str(len(scenes)) + " Done..")
  155. print("------------------------")
  156. time.sleep(10)
  157. if __name__== "__main__":
  158. main()