-
Notifications
You must be signed in to change notification settings - Fork 0
/
Main.py
167 lines (113 loc) · 6.12 KB
/
Main.py
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
import numpy as np
import tkinter as tk
from PIL import Image, ImageTk, ImageGrab
from MNIST_NeuralNet import NetWrapper, MNISTData, Net
class App:
def __init__(self):
master = tk.Tk()
self.init_ui(master)
self.MNIST = MNISTData()
self.NN = NetWrapper()
self.NN.load('Era2')
self.NN.net.eval()
self.input_28 = None
master.mainloop()
def init_ui(self, master):
master.title("Numbr Readr")
master.geometry("900x500")
master.resizable(0, 0)
outline = tk.Frame(master, bg='#6A6A6A')
outline.pack()
user_canvas = DrawingCanvas(outline)
nn_frame = tk.Frame(outline, bg='#3B3E3F', width=350, height=150)
data_frame = tk.Frame(outline, bg='#1F1F1F', width=350, height=120)
buttons_frame = tk.Frame(outline, bg='#1F1F1F', width=350, height=40)
user_canvas.grid(column=0, rowspan=3, padx=50, pady=50)
nn_frame.grid(row=0, column=1, padx=(0, 50), pady=(50, 0))
data_frame.grid(row=1, column=1, padx=(0, 50), pady=(0, 90))
data_frame.grid_propagate(0)
buttons_frame.grid(row=2, column=1, padx=(0, 50), pady=(0, 50), sticky=tk.W)
buttons_frame.grid_propagate(0)
input_bg = tk.PhotoImage(file='BlankOutput.pgm')
input_box = tk.Label(nn_frame, bg='#000000', width=110, height=110, image=input_bg, bd=0)
input_box.photo = input_bg
output_box = tk.Label(nn_frame, bg='#000000', width=110, height=110, image=input_bg, bd=0)
output_box.photo = input_bg
output_box.grid_propagate(0)
input_box.grid(row=0, column=0, padx=(20, 45), pady=20)
output_box.grid(row=0, column=1, padx=(45, 20), pady=20)
prob_frame_list = []
prob_text_list = []
for n in list((0, 1, 2, 3, 4, 5, 6, 7, 8, 9)):
prob_frame_list.append(tk.Frame(data_frame, bg='#1F1F1F', width=80, height=24))
prob_frame_list[n].grid(row=n % 5, column=n // 5, padx=(20,20))
prob_frame_list[n].grid_propagate(0)
prob_text_list.append(tk.Label(prob_frame_list[n], text="", fg="white", bg='#1F1F1F'))
prob_text_list[n].grid(row=0, column=0)
run_button = tk.Button(buttons_frame, text='RUN', command=lambda: self.run_nn(user_canvas, input_box, output_box, prob_text_list), bg='#3C3C3C', fg='#BDBDBD')
run_button.pack(side=tk.LEFT, padx=(0, 10))
clear_button = tk.Button(buttons_frame, text='CLEAR', command=lambda: user_canvas.clear_canvas(input_box), bg='#3C3C3C', fg='#BDBDBD')
clear_button.pack(side=tk.LEFT, padx=(10, 50))
save_button = tk.Button(buttons_frame, text='SAVE', command=lambda: self.save_nn(), bg='#3C3C3C', fg='#BDBDBD')
save_button.pack(side=tk.LEFT, padx=(10, 10))
learn_button = tk.Button(buttons_frame, text='LEARN', command=lambda: self.learn_nn(learn_field), bg='#3C3C3C', fg='#BDBDBD')
learn_button.pack(side=tk.LEFT, padx=(10, 10))
learn_field = tk.Entry(buttons_frame, width=10)
learn_field.pack(side=tk.LEFT, padx=(10, 10))
self.init_bind(user_canvas)
def save_nn(self):
self.NN.save("Era3")
def init_bind(self, user_canvas):
user_canvas.bind("<B1-Motion>", user_canvas.draw_event)
user_canvas.bind("<Button-1>", user_canvas.draw_event)
def run_nn(self, user_canvas, input_box, output_box, prob_text_list):
input_array = Image.fromarray(np.array(user_canvas.ink_matrix.astype('uint8')))
self.input_28 = input_array.resize((28, 28), resample=Image.LANCZOS)
nn_input = np.array(self.input_28)
input_110 = self.input_28.resize((110, 110))
input_image = ImageTk.PhotoImage(input_110)
input_box.config(image=input_image)
input_box.photo = input_image
prob_matrix = np.exp(np.array((self.NN.use(nn_input, self.MNIST.transform))))
np.set_printoptions(precision=2, suppress=True)
output_text = tk.Label(output_box, bg='#000000', bd=0, text=f" {np.argmax(prob_matrix)}", fg="white", font=("",75))
output_text.grid(row=0, column=0)
for n in list((0, 1, 2, 3, 4, 5, 6, 7, 8, 9)):
prob_text_list[n].configure(text=f"{n} : {prob_matrix[0, n]*100:.6f}")
prob_text_list[n].grid(row=0, column=0)
def learn_nn(self, learn_field):
learn_input = int(learn_field.get())
if (learn_input > 9) or (learn_input < 0):
print("Invalid entry")
pass
self.NN.user_learn(np.array(self.input_28), learn_input)
class DrawingCanvas(tk.Canvas):
def __init__(self, outline):
tk.Canvas.__init__(self, outline)
self.config(bg='#E4E4E4', width=400, height=400, cursor='dot')
self.ink_matrix = np.zeros((400, 400))
self.img = tk.PhotoImage(file="BlankCanvas.pgm")
self.create_image((200, 200), image=self.img, state="normal")
def draw_event(self, event):
self.focus_set()
self.gaussian_pen(event.x, event.y)
def clear_canvas(self, input_box):
self.ink_matrix = np.zeros((400, 400))
self.img = tk.PhotoImage(file="BlankCanvas.pgm")
self.create_image((200, 200), image=self.img, state="normal")
def gaussian_pen(self, center_x, center_y):
for x in range(0, 20):
for y in range(0, 20):
ink_dark = 255 - int(400 * np.exp(-1 * ((x ** 2) / (2 * (15 ** 2)) + (y ** 2) / (2 * (15 ** 2)))))
for (pos_x, pos_y) in list(((x,y), (-x,y), (-x,-y), (x,-y))):
if 0 <= center_x + pos_x < 400 and 0 <= center_y + pos_y < 400:
quad1 = self.img.get(center_x + pos_x, center_y + pos_y)
if ink_dark < 0:
ink_dark = 0
if ink_dark > 75:
pass
elif ink_dark < quad1[0]:
ink_hex = '#%02x%02x%02x' % (ink_dark, ink_dark, ink_dark)
self.img.put(ink_hex, (center_x + pos_x, center_y + pos_y))
self.ink_matrix[center_y + pos_y, center_x + pos_x ] = 255 - ink_dark
app = App()