def main():
'''
Parse command line arguments and execute the code
'''
parser = argparse.ArgumentParser()
parser.add_argument('--dataset_path', required=True, type=str)
parser.add_argument('--newext', default='.PNG', type=str)
parser.add_argument('--oldext', default='.JPEG', type=str)
args = parser.parse_args()
start = time.time()
image_list= Utils_Image.get_Image_List(args.dataset_path, args.oldext)
progress = progressbar.ProgressBar(widgets=[progressbar.Bar('=', '[', ']'), ' ',progressbar.Percentage(), ' ',progressbar.ETA()])
print "Start Processing... May take a while..."
for image_path in progress(image_list):
Utils_Image.resizeImage(image_path)
Utils_Image.change_extension(image_path,args.oldext,args.newext)
end = time.time()
print("Parsed: %d Image of the Dataset"%(len(image_list)))
print("Elapsed Time:%d Seconds"%(end-start))
print("Running Completed with Success!!!")
def main():
'''
Parse command line arguments and execute the code
'''
######### TENSORBOX PARAMETERS
start = time.time()
parser = argparse.ArgumentParser()
# parser.add_argument('--result_folder', default='summary_result/', type=str)
# parser.add_argument('--summary_file', default='results.txt', type=str)
parser.add_argument('--output_name', default='output.mp4', type=str)
parser.add_argument('--hypes',
default='./TENSORBOX/hypes/overfeat_rezoom.json',
type=str)
parser.add_argument('--weights',
default='./TENSORBOX/data/save.ckpt-1090000',
type=str)
parser.add_argument('--perc', default=2, type=int)
parser.add_argument('--path_video', required=True, type=str)
args = parser.parse_args()
# hypes_file = './hypes/overfeat_rezoom.json'
# weights_file= './output/save.ckpt-1090000'
path_video_folder = os.path.splitext(os.path.basename(args.path_video))[0]
pred_idl = './%s/%s_val.idl' % (path_video_folder, path_video_folder)
idl_filename = path_video_folder + '/' + path_video_folder + '.idl'
frame_list = []
frame_list = Utils_Video.extract_idl_from_frames(args.path_video,
args.perc,
path_video_folder,
'frames/', idl_filename)
progress = progressbar.ProgressBar(widgets=[
progressbar.Bar('=', '[', ']'), ' ',
progressbar.Percentage(), ' ',
progressbar.ETA()
])
for image_path in progress(frame_list):
Utils_Image.resizeImage(image_path)
Utils_Image.resizeImage(-1)
det_frame_list = Utils_Tensorbox.still_image_TENSORBOX_singleclass(
frame_list, path_video_folder, args.hypes, args.weights, pred_idl)
Utils_Video.make_video_from_list(args.output_name, det_frame_list)
end = time.time()
print("Elapsed Time:%d Seconds" % (end - start))
print("Running Completed with Success!!!")
def main():
'''
Parse command line arguments and execute the code
'''
######### TENSORBOX PARAMETERS
start = time.time()
parser = argparse.ArgumentParser()
parser.add_argument('--det_frames_folder', default='det_frames/', type=str)
parser.add_argument('--det_result_folder', default='det_results/', type=str)
parser.add_argument('--frames_folder', default='frames/', type=str)
# parser.add_argument('--result_folder', default='summary_result/', type=str)
# parser.add_argument('--summary_file', default='results.txt', type=str)
parser.add_argument('--output_name', default='output.mp4', type=str)
parser.add_argument('--hypes', default='./TENSORBOX/hypes/overfeat_rezoom.json', type=str)
parser.add_argument('--weights', default='./TENSORBOX/output/save.ckpt-1090000', type=str)
parser.add_argument('--perc', default=2, type=int)
parser.add_argument('--path_video', required=True, type=str)
args = parser.parse_args()
# hypes_file = './hypes/overfeat_rezoom.json'
# weights_file= './output/save.ckpt-1090000'
path_video_folder = os.path.splitext(os.path.basename(args.path_video))[0]
pred_idl = './%s/%s_val.idl' % (path_video_folder, path_video_folder)
idl_filename=path_video_folder+'/'+path_video_folder+'.idl'
frame_list=[]
frame_list = Utils_Video.extract_idl_from_frames(args.path_video, args.perc, path_video_folder, args.frames_folder, idl_filename )
progress = progressbar.ProgressBar(widgets=[progressbar.Bar('=', '[', ']'), ' ',progressbar.Percentage(), ' ',progressbar.ETA()])
for image_path in progress(frame_list):
Utils_Image.resizeImage(image_path)
det_frame_list=still_image_TENSORBOX(idl_filename, frame_list, args.det_frames_folder, args.det_result_folder, args.frames_folder, path_video_folder, args.hypes, args.weights, pred_idl)
Utils_Video.make_video_from_list(args.output_name, det_frame_list)
end = time.time()
print("Elapsed Time:%d Seconds"%(end-start))
print("Running Completed with Success!!!")
def make_video_from_list(out_vid_path, frames_list):
if frames_list[0] is not None:
img = cv2.imread(frames_list[0], True)
print frames_list[0]
h, w = img.shape[:2]
fourcc = cv2.cv.CV_FOURCC('m', 'p', '4', 'v')
out = cv2.VideoWriter(out_vid_path,fourcc, 20.0, (w, h), True)
print("Start Making File Video:%s " % out_vid_path)
print("%d Frames to Compress"%len(frames_list))
progress = progressbar.ProgressBar(widgets=[progressbar.Bar('=', '[', ']'), ' ',progressbar.Percentage(), ' ',progressbar.ETA()])
for i in progress(range(0,len(frames_list))):
if Utils_Image.check_image_with_pil(frames_list[i]):
out.write(img)
img = cv2.imread(frames_list[i], True)
out.release()
print("Finished Making File Video:%s " % out_vid_path)
def make_video_from_list(out_vid_path, frames_list):
if frames_list[0] is not None:
img = cv2.imread(frames_list[0], True)
print frames_list[0]
h, w = img.shape[:2]
fourcc = cv2.cv.CV_FOURCC('m', 'p', '4', 'v')
out = cv2.VideoWriter(out_vid_path, fourcc, 20.0, (w, h), True)
print("Start Making File Video:%s " % out_vid_path)
print("%d Frames to Compress" % len(frames_list))
progress = progressbar.ProgressBar(widgets=[
progressbar.Bar('=', '[', ']'), ' ',
progressbar.Percentage(), ' ',
progressbar.ETA()
])
for i in progress(range(0, len(frames_list))):
if Utils_Image.check_image_with_pil(frames_list[i]):
out.write(img)
img = cv2.imread(frames_list[i], True)
out.release()
print("Finished Making File Video:%s " % out_vid_path)
def make_tracked_video(out_vid_path, labeled_video_frames):
if labeled_video_frames[0] is not None:
img = cv2.imread(labeled_video_frames[0], True)
print "Reading Filename: %s" % labeled_video_frames[0]
h, w = img.shape[:2]
print "Video Size: width: %d height: %d" % (h, w)
fourcc = cv2.VideoWriter_fourcc('m', 'p', '4', 'v')
out = cv2.VideoWriter(out_vid_path, fourcc, 20.0, (w, h), True)
print("Start Making File Video:%s " % out_vid_path)
print("%d Frames to Compress" % len(labeled_video_frames))
progress = progressbar.ProgressBar(widgets=[
progressbar.Bar('=', '[', ']'), ' ',
progressbar.Percentage(), ' ',
progressbar.ETA()
])
for i in progress(range(0, len(labeled_video_frames))):
if Utils_Image.check_image_with_pil(labeled_video_frames[i]):
out.write(img)
img = cv2.imread(labeled_video_frames[i], True)
out.release()
print("Finished Making File Video:%s " % out_vid_path)
def parse_XML_to_multiclass_txt(bb_XML_file_list, path_val_folder, path_dataset):
count_rect=0
count_img=0
progress = progressbar.ProgressBar(widgets=[progressbar.Bar('=', '[', ']'), ' ',progressbar.Percentage(), ' ',progressbar.ETA()])
print "Start Processing & Building the MultiClass Dataset... May take a while..."
path_mltcl_bb_file=path_dataset+'/'+string_mltcl_bb_file # Create this file in .dataset/airplane/airplane_bb_mltcl_file_list.txt
path_mltcl_class_code_file=path_dataset+'/'+string_mltcl_class_code_file
path_mltcl_class_name_file=path_dataset+'/'+string_mltcl_class_name_file
path_mltcl_chall_code_file=path_dataset+'/'+string_mltcl_chall_code_file
for file_name in progress(bb_XML_file_list):
with open(file_name, 'rt') as f:
tree = ElementTree.parse(f)
for obj in tree.findall('object'):
name = obj.find('name').text
class_code= name
name = Classes.code_to_class_string(name)
if name in ["nothing"]:
continue
else:
path_orig_file=path_val_folder
jump=0
image_multi_class= Picture_Info()
image_multi_class.dataset_path= path_val_folder
rectangle_multi= BB_Rectangle()
#xmin x1 letf
#ymin y1 bottom
#xmax x2 right
#ymax y2 top
for node in tree.iter():
tag=str(node.tag)
if tag in ["folder"]:
path_orig_file=path_orig_file+'/'+str(node.text)
image_multi_class.folder= str(node.text)
if tag in ["filename"]:
image_multi_class.filename=str(node.text)+'.PNG'
path_orig_file=path_orig_file+'/'+str(node.text)+'.JPEG'
if tag in ["width"]:
image_multi_class.width= int(node.text)
if tag in ["height"]:
image_multi_class.height= int(node.text)
if tag in ["trackid"]:
image_multi_class.frame= int(os.path.basename(file_name).replace('.xml', ''))
if tag in ['name']:
if str(Classes.code_to_class_string(str(node.text))) in ["nothing"]:
jump = 1
else :
jump=0
rectangle_multi.label_chall=int(Classes.class_string_to_comp_code(str(Classes.code_to_class_string(str(node.text)))))
rectangle_multi.label_code=str(node.text)
rectangle_multi.label=Classes.code_to_class_string(str(node.text))
if tag in ["xmax"]:
if jump == 0:
rectangle_multi.x2=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), True))
if tag in ["xmin"]:
if jump == 0:
rectangle_multi.x1=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), True))
if tag in ["ymin"]:
if jump == 0:
rectangle_multi.y2=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), False))
if tag in ["ymax"]:
if jump == 0:
rectangle_multi.y1=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), False))
image_multi_class.append_rect(rectangle_multi)
count_rect=count_rect+1
if jump == 0:
count_img=count_img+1
out_stream = open(path_mltcl_bb_file, "a")
out_stream.write(image_multi_class.get_info_string()+ os.linesep)
out_stream = open(path_mltcl_class_code_file, "a")
out_stream.write(image_multi_class.get_rects_code()+ os.linesep)
out_stream = open(path_mltcl_class_name_file, "a")
out_stream.write(image_multi_class.get_rects_labels()+ os.linesep)
out_stream = open(path_mltcl_chall_code_file, "a")
out_stream.write(image_multi_class.get_rects_chall() + os.linesep)
break
out_stream = open('dataset/summary_multiclass.txt', "w")
out_stream.write( "SUMMARY:" + os.linesep+ ("Parsed: %d BB Files"%len(bb_XML_file_list)) + os.linesep+ ("Added: %d Object Rectangles"%count_rect) + os.linesep+("Added: %d Images"%count_img)+os.linesep)
print "Ended with Success the Process"
print "SUMMARY:"
print "Parsed: %d BB Files"%len(bb_XML_file_list)
print "Added: %d Object Rectangles"%count_rect
print "Added: %d Images"%count_img
if count_img>0:
print "Ratio BB_Files/Images: %.2f"%(float(count_img)/float(len(bb_XML_file_list)))
out_stream.write(("Ratio BB_Files/Images: %.2f"%(float(count_img)/float(len(bb_XML_file_list))))+os.linesep)
if count_rect>0:
print "Ratio Images/Object_Rectangles: %.2f"%(float(count_img)/float(count_rect))
out_stream.write(("Ratio Images/Object_Rectangles: %.2f"%(float(count_img)/float(count_rect)))+os.linesep)
def still_image_TENSORBOX_multiclass(frames_list,path_video_folder,hypes_file,weights_file,pred_idl):
from train import build_forward
print("Starting DET Phase")
if not os.path.exists(path_video_folder+'/'+folder_path_det_frames):
os.makedirs(path_video_folder+'/'+folder_path_det_frames)
print("Created Folder: %s"%path_video_folder+'/'+folder_path_det_frames)
if not os.path.exists(path_video_folder+'/'+folder_path_det_result):
os.makedirs(path_video_folder+'/'+folder_path_det_result)
print("Created Folder: %s"% path_video_folder+'/'+folder_path_det_result)
det_frames_list=[]
#### START TENSORBOX CODE ###
idl_filename=path_video_folder+'/'+path_video_folder+'.idl'
### Opening Hypes file for parameters
with open(hypes_file, 'r') as f:
H = json.load(f)
### Building Network
tf.reset_default_graph()
googlenet = googlenet_load.init(H)
x_in = tf.placeholder(tf.float32, name='x_in', shape=[H['image_height'], H['image_width'], 3])
if H['use_rezoom']:
pred_boxes, pred_logits, pred_confidences, pred_confs_deltas, pred_boxes_deltas = build_forward(H, tf.expand_dims(x_in, 0), googlenet, 'test', reuse=None)
grid_area = H['grid_height'] * H['grid_width']
pred_confidences = tf.reshape(tf.nn.softmax(tf.reshape(pred_confs_deltas, [grid_area * H['rnn_len'], H['num_classes']])), [grid_area, H['rnn_len'], H['num_classes']])
pred_logits = tf.reshape(tf.nn.softmax(tf.reshape(pred_logits, [grid_area * H['rnn_len'], H['num_classes']])), [grid_area, H['rnn_len'], H['num_classes']])
if H['reregress']:
pred_boxes = pred_boxes + pred_boxes_deltas
else:
pred_boxes, pred_logits, pred_confidences = build_forward(H, tf.expand_dims(x_in, 0), googlenet, 'test', reuse=None)
saver = tf.train.Saver()
with tf.Session() as sess:
sess.run(tf.initialize_all_variables())
saver.restore(sess, weights_file )##### Restore a Session of the Model to get weights and everything working
annolist = al.AnnoList()
#### Starting Evaluating the images
lenght=int(len(frames_list))
print("%d Frames to DET"%len(frames_list))
progress = progressbar.ProgressBar(widgets=[progressbar.Bar('=', '[', ']'), ' ',progressbar.Percentage(), ' ',progressbar.ETA()])
frameNr=0
skipped=0
for i in progress(range(0, len(frames_list))):
if Utils_Image.isnotBlack(frames_list[i]) & Utils_Image.check_image_with_pil(frames_list[i]):
img = imread(frames_list[i])
feed = {x_in: img}
(np_pred_boxes,np_pred_logits, np_pred_confidences) = sess.run([pred_boxes,pred_logits, pred_confidences], feed_dict=feed)
# print_logits(np_pred_confidences)
pred_anno = al.Annotation()
#pred_anno.imageName = test_anno.imageName
# print "np_pred_confidences shape" + str(np_pred_confidences.shape)
# print "np_pred_boxes shape" + str(np_pred_boxes.shape)
# for i in range(0, np_pred_confidences.shape[0]):
# print np_pred_confidences[i]
# for j in range(0, np_pred_confidences.shape[2]):
# print np_pred_confidences[i][0][j]
rects = get_multiclass_rectangles(H, np_pred_confidences, np_pred_boxes, rnn_len=H['rnn_len'])
pred_anno.rects = rects
pred_anno.imageName = frames_list[i]
pred_anno.frameNr = frameNr
frameNr=frameNr+1
det_frames_list.append(frames_list[i])
pick = NMS(rects)
draw_rectangles(frames_list[i],frames_list[i], pick)
annolist.append(pred_anno)
else: skipped=skipped+1
saveTextResults(idl_filename,annolist)
annolist.save(pred_idl)
print("Skipped %d Black Frames"%skipped)
#### END TENSORBOX CODE ###
return det_frames_list
def main():
'''
Parse command line arguments and execute the code
'''
######### TENSORBOX PARAMETERS
start = time.time()
parser = argparse.ArgumentParser()
# parser.add_argument('--result_folder', default='summary_result/', type=str)
# parser.add_argument('--summary_file', default='results.txt', type=str)
parser.add_argument('--output_name', default='output.mp4', type=str)
parser.add_argument('--hypes',
default='./TENSORBOX/hypes/overfeat_rezoom.json',
type=str)
parser.add_argument('--weights',
default='./TENSORBOX/data/save.ckpt-1250000',
type=str)
parser.add_argument('--perc', default=100, type=int)
parser.add_argument('--path_video',
default='DAVIS-2017-TrainVal.mp4',
type=str) # required=True, type=str)
args = parser.parse_args()
# hypes_file = './hypes/overfeat_rezoom.json'
# weights_file= './output/save.ckpt-1090000'
path_video_folder = os.path.splitext(os.path.basename(args.path_video))[0]
pred_idl = './%s/%s_val.idl' % (path_video_folder, path_video_folder)
idl_filename = path_video_folder + '/' + path_video_folder + '.idl'
frame_tensorbox = []
frame_inception = []
frame_tensorbox, frame_inception = Utils_Video.extract_frames_incten(
args.path_video, args.perc, path_video_folder, idl_filename)
progress = progressbar.ProgressBar(widgets=[
progressbar.Bar('=', '[', ']'), ' ',
progressbar.Percentage(), ' ',
progressbar.ETA()
])
for image_path in progress(frame_tensorbox):
Utils_Image.resizeImage(image_path)
Utils_Image.resizeImage(-1)
video_info = Utils_Tensorbox.bbox_det_TENSORBOX_multiclass(
frame_tensorbox, path_video_folder, args.hypes, args.weights, pred_idl)
tracked_video = Utils_Video.recurrent_track_objects(video_info)
# tracked_video=utils_video.track_objects(video_info)
# labeled_video=Utils_Imagenet.label_video(tracked_video, frame_inception)
labeled_video = Utils_Imagenet.recurrent_label_video(
tracked_video, frame_inception)
# tracked_video=utils_video.track_objects(video_info)
# tracked_video=utils_video.track_and_label_objects(video_info)
labeled_frames = Utils_Video.draw_rectangles(path_video_folder,
labeled_video)
Utils_Video.make_tracked_video(args.output_name, labeled_frames)
frame.saveVideoResults(idl_filename, labeled_video)
# utils_video.make_tracked_video(args.output_name, labeled_video)
end = time.time()
print("Elapsed Time:%d Seconds" % (end - start))
print("Running Completed with Success!!!")
def parse_XML_lightweight_txt(bb_XML_file_list, path_val_folder, path_dataset):
count_rect = 0
count_img = 0
progress = progressbar.ProgressBar(widgets=[progressbar.Bar('=', '[', ']'), ' ',progressbar.Percentage(), ' ',progressbar.ETA()])
print "Start Processing & Building Dataset... may take a while..."
path_mltcl_bb_file=path_dataset+'/'+string_mltcl_bb_file # Create this file in .dataset/airplane/airplane_bb_mltcl_file_list.txt
path_mltcl_class_code_file=path_dataset+'/'+string_mltcl_class_code_file
path_mltcl_class_name_file=path_dataset+'/'+string_mltcl_class_name_file
path_mltcl_chall_code_file=path_dataset+'/'+string_mltcl_chall_code_file
for file_name in progress(bb_XML_file_list):
with open(file_name, 'rt') as f:
tree = ElementTree.parse(f)
for obj in tree.findall('object'):
name = obj.find('name').text
class_code= name
name = Classes.code_to_class_string(name)
if name in ["nothing"]:
continue
else:
same_label=0
#The files with the original data path are made in both: multiclass e single class
path_bb_file=path_dataset+'/'+name+'/'+ name+string_bb_file
path_class_code_file= path_dataset+'/'+name+'/'+name+string_class_code_file
path_class_name_file= path_dataset+'/'+name+'/'+name+string_class_name_file
path_chall_code_file= path_dataset+'/'+name+'/'+name+string_chall_code_file
path_orig_file=path_val_folder
jump=0
image_single_class= Picture_Info()
image_single_class.dataset_path= path_val_folder
image_multi_class= Picture_Info()
image_multi_class.dataset_path= path_val_folder
rectangle_single= BB_Rectangle()
rectangle_multi= BB_Rectangle()
#xmin x1 letf
#ymin y1 bottom
#xmax x2 right
#ymax y2 top
for node in tree.iter():
tag=str(node.tag)
if tag in ["folder"]:
path_orig_file=path_orig_file+'/'+str(node.text)
image_single_class.folder= str(node.text)
image_multi_class.folder= str(node.text)
if tag in ["filename"]:
image_single_class.filename=str(node.text)+'.PNG'
image_multi_class.filename=str(node.text)+'.PNG'
path_orig_file=path_orig_file+'/'+str(node.text)+'.JPEG'
if tag in ['name']:
if str(Classes.code_to_class_string(str(node.text))) in ["nothing"]:
jump = 1
else :
jump=0
rectangle_multi.label_chall=int(Classes.class_string_to_comp_code(str(Classes.code_to_class_string(str(node.text)))))
rectangle_multi.label_code=str(node.text)
rectangle_multi.label=Classes.code_to_class_string(str(node.text))
if str(node.text) == class_code:
same_label = 1
rectangle_single.label_chall=int(Classes.class_string_to_comp_code(str(Classes.code_to_class_string(str(node.text)))))
rectangle_single.label_code=str(node.text)
rectangle_single.label=Classes.code_to_class_string(str(node.text))
else: same_label = 0
if tag in ["xmax"]:
if jump == 0:
rectangle_multi.x2=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), False))
if same_label==1:
rectangle_single.x2=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), False))
if tag in ["xmin"]:
if jump == 0:
rectangle_multi.x1=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), False))
if same_label==1:
rectangle_single.x1=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), False))
if tag in ["ymax"]:
if jump == 0:
rectangle_multi.y2=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), False))
image_multi_class.append_rect(rectangle_multi)
count_rect=count_rect+1
if same_label==1:
rectangle_single.y2=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), False))
image_single_class.append_rect(rectangle_single)
if tag in ["ymin"]:
if jump == 0:
rectangle_multi.y1=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), False))
if same_label==1:
rectangle_single.y1=float(Utils_Image.transform_point(image_multi_class.width,image_multi_class.height,width, height,float(node.text), False))
if jump == 0:
count_img=count_img+1
out_stream = open(path_mltcl_bb_file, "a")
out_stream.write(image_multi_class.get_info_string()+ os.linesep)
out_stream = open(path_mltcl_class_code_file, "a")
out_stream.write(image_multi_class.get_rects_code()+ os.linesep)
out_stream = open(path_mltcl_class_name_file, "a")
out_stream.write(image_multi_class.get_rects_labels()+ os.linesep)
out_stream = open(path_mltcl_chall_code_file, "a")
out_stream.write(image_multi_class.get_rects_chall() + os.linesep)
if same_label==1:
out_stream = open(path_bb_file, "a")
out_stream.write(image_single_class.get_info_string()+ os.linesep)
out_stream = open(path_class_code_file, "a")
out_stream.write(image_single_class.get_rects_chall()+ os.linesep)
out_stream = open(path_class_name_file, "a")
out_stream.write(image_single_class.get_rects_labels()+ os.linesep)
out_stream = open(path_chall_code_file, "a")
out_stream.write(image_single_class.get_rects_code() + os.linesep)
break
print "SUMMARY:"
print "Parsed: %d BB Files"%len(bb_XML_file_list)
print "Added: %d Object Rectangles"%count_rect
print "Added: %d Images"%count_img
if count_img>0:
print "Ratio BB_Files/Images: %.2f"%(float(count_img)/float(len(bb_XML_file_list)))
if count_rect>0:
print "Ratio Images/Object_Rectangles: %.2f"%(float(count_img)/float(count_rect))