machine based reading order inference & optimized algorithm

machine_based_reading_order_integration
vahidrezanezhad 7 months ago
parent 49c93149a4
commit 59c0d90e5a

@ -2855,7 +2855,6 @@ class Eynollah:
model = load_model(model_file , compile=False,custom_objects = {"PatchEncoder": PatchEncoder, "Patches": Patches})
return model
def do_order_of_regions_with_machine(self,contours_only_text_parent, contours_only_text_parent_h, text_regions_p):
y_len = text_regions_p.shape[0]
x_len = text_regions_p.shape[1]
@ -2983,6 +2982,154 @@ class Eynollah:
id_of_texts.append( REGION_ID_TEMPLATE % order )
return order_of_texts, id_of_texts
def update_list_and_return_first_biger_than_one_length(self,index_element_to_be_updated, innner_index_pr_pos, pr_list, pos_list,list_inp):
list_inp.pop(index_element_to_be_updated)
if len(pr_list)>0:
list_inp.insert(index_element_to_be_updated, pr_list)
else:
index_element_to_be_updated = index_element_to_be_updated -1
list_inp.insert(index_element_to_be_updated+1, [innner_index_pr_pos])
if len(pos_list)>0:
list_inp.insert(index_element_to_be_updated+2, pos_list)
len_all_elements = [len(i) for i in list_inp]
list_len_bigger_1 = np.where(np.array(len_all_elements)>1)
list_len_bigger_1 = list_len_bigger_1[0]
if len(list_len_bigger_1)>0:
early_list_bigger_than_one = list_len_bigger_1[0]
else:
early_list_bigger_than_one = -20
return list_inp, early_list_bigger_than_one
def do_order_of_regions_with_machine_optimized_algorithm(self,contours_only_text_parent, contours_only_text_parent_h, text_regions_p):
y_len = text_regions_p.shape[0]
x_len = text_regions_p.shape[1]
img_poly = np.zeros((y_len,x_len), dtype='uint8')
unique_pix = np.unique(text_regions_p)
img_poly[text_regions_p[:,:]==1] = 1
img_poly[text_regions_p[:,:]==2] = 2
img_poly[text_regions_p[:,:]==3] = 4
img_poly[text_regions_p[:,:]==6] = 5
model_ro_machine, _ = self.start_new_session_and_model(self.model_reading_order_machine_dir)
height1 =672#448
width1 = 448#224
height2 =672#448
width2= 448#224
height3 =672#448
width3 = 448#224
_, cy_main, x_min_main, x_max_main, y_min_main, y_max_main, _ = find_new_features_of_contours(contours_only_text_parent_h)
img_header_and_sep = np.zeros((y_len,x_len), dtype='uint8')
for j in range(len(cy_main)):
img_header_and_sep[int(y_max_main[j]):int(y_max_main[j])+12,int(x_min_main[j]):int(x_max_main[j]) ] = 1
co_text_all = contours_only_text_parent + contours_only_text_parent_h
labels_con = np.zeros((y_len,x_len,len(co_text_all)),dtype='uint8')
for i in range(len(co_text_all)):
img_label = np.zeros((y_len,x_len,3),dtype='uint8')
img_label=cv2.fillPoly(img_label, pts =[co_text_all[i]], color=(1,1,1))
labels_con[:,:,i] = img_label[:,:,0]
img3= np.copy(img_poly)
labels_con = resize_image(labels_con, height1, width1)
img_header_and_sep = resize_image(img_header_and_sep, height1, width1)
img3= resize_image (img3, height3, width3)
img3 = img3.astype(np.uint16)
inference_bs = 4
input_1= np.zeros( (inference_bs, height1, width1,3))
starting_list_of_regions = []
starting_list_of_regions.append( list(range(labels_con.shape[2])) )
index_update = 0
index_selected = starting_list_of_regions[0]
#print(labels_con.shape[2],"number of regions for reading order")
while index_update>=0:
ij_list = starting_list_of_regions[index_update]
i = ij_list[0]
ij_list.pop(0)
pr_list = []
post_list = []
batch_counter = 0
tot_counter = 1
tot_iteration = len(ij_list)
full_bs_ite= tot_iteration//inference_bs
last_bs = tot_iteration % inference_bs
jbatch_indexer =[]
for j in ij_list:
img1= np.repeat(labels_con[:,:,i][:, :, np.newaxis], 3, axis=2)
img2 = np.repeat(labels_con[:,:,j][:, :, np.newaxis], 3, axis=2)
img2[:,:,0][img3[:,:]==5] = 2
img2[:,:,0][img_header_and_sep[:,:]==1] = 3
img1[:,:,0][img3[:,:]==5] = 2
img1[:,:,0][img_header_and_sep[:,:]==1] = 3
jbatch_indexer.append(j)
input_1[batch_counter,:,:,0] = img1[:,:,0]/3.
input_1[batch_counter,:,:,2] = img2[:,:,0]/3.
input_1[batch_counter,:,:,1] = img3[:,:]/5.
batch_counter = batch_counter+1
if batch_counter==inference_bs or ( (tot_counter//inference_bs)==full_bs_ite and tot_counter%inference_bs==last_bs):
y_pr=model_ro_machine.predict(input_1 , verbose=0)
if batch_counter==inference_bs:
iteration_batches = inference_bs
else:
iteration_batches = last_bs
for jb in range(iteration_batches):
if y_pr[jb][0]>=0.5:
post_list.append(jbatch_indexer[jb])
else:
pr_list.append(jbatch_indexer[jb])
batch_counter = 0
jbatch_indexer = []
tot_counter = tot_counter+1
starting_list_of_regions, index_update = self.update_list_and_return_first_biger_than_one_length(index_update, i, pr_list, post_list,starting_list_of_regions)
index_sort = [i[0] for i in starting_list_of_regions ]
REGION_ID_TEMPLATE = 'region_%04d'
order_of_texts = []
id_of_texts = []
for order, id_text in enumerate(index_sort):
order_of_texts.append(id_text)
id_of_texts.append( REGION_ID_TEMPLATE % order )
return order_of_texts, id_of_texts
def run(self):
@ -3252,7 +3399,7 @@ class Eynollah:
if self.full_layout:
if self.reading_order_machine_based:
order_text_new, id_of_texts_tot = self.do_order_of_regions_with_machine(contours_only_text_parent, contours_only_text_parent_h, text_regions_p)
order_text_new, id_of_texts_tot = self.do_order_of_regions_with_machine_optimized_algorithm(contours_only_text_parent, contours_only_text_parent_h, text_regions_p)
else:
if np.abs(slope_deskew) < SLOPE_THRESHOLD:
order_text_new, id_of_texts_tot = self.do_order_of_regions(contours_only_text_parent, contours_only_text_parent_h, boxes, textline_mask_tot)
@ -3268,7 +3415,7 @@ class Eynollah:
else:
contours_only_text_parent_h = None
if self.reading_order_machine_based:
order_text_new, id_of_texts_tot = self.do_order_of_regions_with_machine(contours_only_text_parent, contours_only_text_parent_h, text_regions_p)
order_text_new, id_of_texts_tot = self.do_order_of_regions_with_machine_optimized_algorithm(contours_only_text_parent, contours_only_text_parent_h, text_regions_p)
else:
if np.abs(slope_deskew) < SLOPE_THRESHOLD:
order_text_new, id_of_texts_tot = self.do_order_of_regions(contours_only_text_parent, contours_only_text_parent_h, boxes, textline_mask_tot)

Loading…
Cancel
Save