1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
|
"""!
@brief Example of application for digit recognition based on self-organized feature map.
Digits for 0 to 9 can be recognized. The application has GUI that provides following
function: learning, drawing, recognition, dump saving/loading.
@authors Andrei Novikov (pyclustering@yandex.ru)
@date 2014-2020
@copyright BSD-3-Clause
"""
from pyclustering.nnet.som import som, type_conn;
from pyclustering.samples.definitions import IMAGE_DIGIT_SAMPLES;
from pyclustering.utils import read_image, rgb2gray;
from tkinter import *;
from tkinter import messagebox;
import math;
import pickle;
import os;
import random;
class recognizer:
__network = None;
def __init__(self):
self.__decode_map = [];
for index_digit in range(0, 10, 1):
list_file_digit_sample = IMAGE_DIGIT_SAMPLES.GET_LIST_IMAGE_SAMPLES(index_digit);
for file_name in list_file_digit_sample:
self.__decode_map.append(index_digit);
def train(self):
samples = [];
print("Digit images preprocessing...");
for index_digit in range(0, 10, 1):
list_file_digit_sample = IMAGE_DIGIT_SAMPLES.GET_LIST_IMAGE_SAMPLES(index_digit);
for file_name in list_file_digit_sample:
data = read_image(file_name);
image_pattern = rgb2gray(data);
for index_pixel in range(len(image_pattern)):
if (image_pattern[index_pixel] < 128):
image_pattern[index_pixel] = 1;
else:
image_pattern[index_pixel] = 0;
samples += [ image_pattern ];
print("SOM initialization...");
self.__network = som(2, 5, type_conn.grid_four, None, True);
print("SOM training...");
self.__network.train(samples, 300);
print("SOM is ready...");
def recognize(self, input_pattern):
index_neuron = self.__network.simulate(input_pattern);
decoded_capture_objects = [];
for index_capture_object in self.__network.capture_objects[index_neuron]:
# print("\t%s" % decode_map[index_capture_object]);
decoded_capture_objects.append(self.__decode_map[index_capture_object]);
frequent_index = max(set(decoded_capture_objects), key = decoded_capture_objects.count);
print(decoded_capture_objects);
return frequent_index;
def save_knowledge(self):
result_saving = False;
if (self.__network is not None):
file_network_dump = open("knowledge_recognition_memory_dump", "wb");
pickle.dump(self.__network, file_network_dump);
result_saving = True;
return result_saving;
def load_knowledge(self):
result_loading = False;
if (os.path.isfile("knowledge_recognition_memory_dump") is True):
file_network_dump = open("knowledge_recognition_memory_dump", "rb");
self.__network = pickle.load(file_network_dump);
result_loading = True;
return result_loading;
class digit_application:
__color = "#000000";
__widget = None;
__user_pattern = None;
__recognizer = None;
__master = None;
def __init__(self):
self.__master = Tk();
self.__master.title("Recognition");
self.__widget = Canvas(self.__master, width = 320, height = 320);
self.__widget.pack(expand = YES, fill = BOTH);
self.__widget.bind("<B1-Motion>", self.__paint);
button_recognize = Button(self.__master, text = "Recognize", command = self.click_recognize, width = 25);
button_recognize.pack(side = BOTTOM);
button_recognize = Button(self.__master, text = "Random Image", command = self.click_image_load, width = 25);
button_recognize.pack(side = BOTTOM);
# button_save = Button(self.__master, text = "Save", command = self.click_save, width = 25);
# button_save.pack(side = BOTTOM);
#
# button_load = Button(self.__master, text = "Load", command = self.click_load, width = 25);
# button_load.pack(side = BOTTOM);
button_train = Button(self.__master, text = "Train", command = self.click_train, width = 25);
button_train.pack(side = BOTTOM);
button_clean = Button(self.__master, text = "Clean", command = self.click_clean, width = 25);
button_clean.pack(side = BOTTOM);
self.__user_pattern = [ 0 for i in range(32 * 32) ];
self.__recognizer = recognizer();
def __paint(self, event):
# calculate square that is belong this click
if ( (event.x >= 0) and (event.x < 320) and (event.y >= 0) and (event.y < 320) ):
x1, y1 = math.floor(event.x / 10), math.floor(event.y / 10);
self.__user_pattern[y1 * 32 + x1] = 1;
index2 = (y1 + 1) * 32 + x1;
index3 = y1 * 32 + (x1 + 1);
index4 = (y1 + 1) * 32 + (x1 + 1);
if (index2 < len(self.__user_pattern)):
self.__user_pattern[index2] = 1;
if (index3 < len(self.__user_pattern)):
self.__user_pattern[index3] = 1;
if (index4 < len(self.__user_pattern)):
self.__user_pattern[index4] = 1;
display_x1, display_y1 = x1 * 10, y1 * 10;
display_x2, display_y2 = display_x1 + 20, display_y1 + 20;
self.__widget.create_rectangle(display_x1, display_y1, display_x2, display_y2, fill = self.__color, width = 0);
def click_train(self):
self.__recognizer.train();
def click_load(self):
if (self.__recognizer.load_knowledge() is not True):
messagebox.showwarning("Recognition - Knowledge Loading", "Knowledge represented by self-organized feature map has not been "
"load from hardware to recognizer due to lack of saved dump of that object. "
"Please save knowledge dump after training and after that it will be possible "
"to use load it at any time.");
def click_save(self):
if (self.__recognizer.save_knowledge() is not True):
messagebox.showwarning("Recognition - Knowledge Saving", "Knowledge represented by self-organized feature map has been created "
"because training has been performed. Please train recognizer and after save result of training.");
def click_recognize(self):
digit_index = self.__recognizer.recognize(self.__user_pattern);
messagebox.showinfo("Recognition - Result", "Most probably input digit is " + str(digit_index));
def click_clean(self):
self.__user_pattern = [ 0 for i in range(32 * 32) ];
Canvas.delete(self.__widget, "all");
def click_image_load(self):
self.__user_pattern = [ 0 for i in range(32 * 32) ];
Canvas.delete(self.__widget, "all");
index_digit = int(math.floor(random.random() * 10));
list_file_digit_sample = IMAGE_DIGIT_SAMPLES.GET_LIST_IMAGE_SAMPLES(index_digit);
index_image = int(math.floor( random.random() * len(list_file_digit_sample) ));
file_name = list_file_digit_sample[index_image];
data = read_image(file_name);
image_pattern = rgb2gray(data);
for y in range(32):
for x in range(32):
linear_index = y * 32 + x;
if (image_pattern[linear_index] < 128):
self.__user_pattern[linear_index] = 1;
self.__widget.create_rectangle(x * 10, y * 10, x * 10 + 10, y * 10 + 10, fill = self.__color, width = 0);
def start(self):
mainloop();
app = digit_application();
app.start();
# digit_recognition();
|
