rice-disease-classifier-2

March 1, 2024

[1]: from __future__ import absolute_import, print_function,division,␣

↪unicode_literals

import os.path
import glob
import shutil

import tensorflow as tf
assert tf.__version__.startswith('2')

from tensorflow import keras

from tensorflow.keras.models import Sequential
from tensorflow.keras.preprocessing.image import ImageDataGenerator
from tensorflow.keras.layers import␣
↪Conv2D,Flatten,MaxPooling2D,Dropout,Dense,Activation

from keras import regularizers

import keras

import numpy as np
import matplotlib.pyplot as plt
import pathlib

print ('successful')

successful

[2]: print(tf.__version__)

2.4.1

[3]: BATCH_SIZE = 8
IMG_HEIGHT = 224
IMG_WIDTH = 224

[4]: data_dir ="../input/rice-leaf-diseases/rice_leaf_diseases"

data_dir = pathlib.Path(data_dir)

CLASS_NAMES = np.array(['Leaf Blight','Brown Spot','Leaf Smut'])

1
 print('Class Names: ', CLASS_NAMES)

Class Names: ['Leaf Blight' 'Brown Spot' 'Leaf Smut']

[5]: train_path = '../input/rice-leaf-diseases/rice_leaf_diseases'

test_path = '../input/rice-leaf-diseases/rice_leaf_diseases'

[6]: image_train_gen = ImageDataGenerator(rescale=1./255,

zoom_range=0.50,
rotation_range=45,
horizontal_flip=True,
width_shift_range=0.15,
height_shift_range=0.15)

train_data_gen = image_train_gen.flow_from_directory(train_path,
shuffle=True,
batch_size=BATCH_SIZE,
␣
↪target_size=(IMG_HEIGHT,IMG_WIDTH),

class_mode='sparse')

img_val_gen = ImageDataGenerator(rescale=1./255)
val_data_gen = img_val_gen.flow_from_directory(test_path,
batch_size=BATCH_SIZE,
␣
↪target_size=(IMG_HEIGHT,IMG_WIDTH),

class_mode='sparse')

Found 120 images belonging to 3 classes.

Found 120 images belonging to 3 classes.

[7]: def plotImages(image_arr):

fig,axes = plt.subplots(1, 5, figsize=(20,20))
axes = axes.flatten()
for img,ax in zip(image_arr,axes):
ax.imshow(img)
plt.tight_layout()
plt.show()

[8]: # Plot a few training images

img_array = [train_data_gen[0][0][0] for i in range(5)]
plotImages(img_array)

2
[9]: # plot a few val images
img_array = [val_data_gen[0][0][0] for i in range(5)]
plotImages(img_array)

[10]: # Model building

#Instatiating A convnet

model = Sequential()
model.add(Conv2D(16, (3,3), input_shape=(224,224,3), activation="relu"))
model.add(MaxPooling2D(pool_size = (2,2)))
model.add(Conv2D(32, (3,3), activation="relu"))
model.add(MaxPooling2D(pool_size = (2,2)))
model.add(Conv2D(64, (3,3), activation="relu"))
model.add(MaxPooling2D(pool_size = (2,2)))
model.add(Flatten())
model.add(Dropout(0.2))
model.add(Dense(128,activation="relu"))
model.add(Dropout(0.2))
model.add(Dense(3, activation="softmax"))

model.compile(
optimizer = "adam",
loss = "sparse_categorical_crossentropy",
metrics = ['accuracy']
)

model.summary()

Model: "sequential"

3
 _________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d (Conv2D) (None, 222, 222, 16) 448
_________________________________________________________________
max_pooling2d (MaxPooling2D) (None, 111, 111, 16) 0
_________________________________________________________________
conv2d_1 (Conv2D) (None, 109, 109, 32) 4640
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 54, 54, 32) 0
_________________________________________________________________
conv2d_2 (Conv2D) (None, 52, 52, 64) 18496
_________________________________________________________________
max_pooling2d_2 (MaxPooling2 (None, 26, 26, 64) 0
_________________________________________________________________
flatten (Flatten) (None, 43264) 0
_________________________________________________________________
dropout (Dropout) (None, 43264) 0
_________________________________________________________________
dense (Dense) (None, 128) 5537920
_________________________________________________________________
dropout_1 (Dropout) (None, 128) 0
_________________________________________________________________
dense_1 (Dense) (None, 3) 387
=================================================================
Total params: 5,561,891
Trainable params: 5,561,891
Non-trainable params: 0
_________________________________________________________________

[11]: EPOCHS=50
history = model.fit_generator(train_data_gen, epochs=EPOCHS,␣
↪validation_data=val_data_gen)

/opt/conda/lib/python3.7/site-
packages/tensorflow/python/keras/engine/training.py:1844: UserWarning:
`Model.fit_generator` is deprecated and will be removed in a future version.
Please use `Model.fit`, which supports generators.
warnings.warn('`Model.fit_generator` is deprecated and '
Epoch 1/50
15/15 [==============================] - 12s 777ms/step - loss: 3.0111 -
accuracy: 0.2745 - val_loss: 1.1084 - val_accuracy: 0.3333
Epoch 2/50
15/15 [==============================] - 10s 654ms/step - loss: 1.1207 -
accuracy: 0.2794 - val_loss: 1.0957 - val_accuracy: 0.4833
Epoch 3/50
15/15 [==============================] - 10s 650ms/step - loss: 1.0968 -

4
accuracy: 0.3379 - val_loss: 1.0924 - val_accuracy: 0.4750
Epoch 4/50
15/15 [==============================] - 9s 639ms/step - loss: 1.0867 -
accuracy: 0.3765 - val_loss: 1.2203 - val_accuracy: 0.3333
Epoch 5/50
15/15 [==============================] - 10s 654ms/step - loss: 1.1741 -
accuracy: 0.3260 - val_loss: 1.0914 - val_accuracy: 0.3333
Epoch 6/50
15/15 [==============================] - 9s 638ms/step - loss: 1.0748 -
accuracy: 0.3696 - val_loss: 1.0638 - val_accuracy: 0.4917
Epoch 7/50
15/15 [==============================] - 10s 657ms/step - loss: 1.0784 -
accuracy: 0.4198 - val_loss: 1.0498 - val_accuracy: 0.5083
Epoch 8/50
15/15 [==============================] - 10s 661ms/step - loss: 1.0618 -
accuracy: 0.3689 - val_loss: 1.0554 - val_accuracy: 0.5167
Epoch 9/50
15/15 [==============================] - 10s 645ms/step - loss: 1.0693 -
accuracy: 0.4649 - val_loss: 1.0337 - val_accuracy: 0.4000
Epoch 10/50
15/15 [==============================] - 9s 639ms/step - loss: 1.0641 -
accuracy: 0.3904 - val_loss: 1.0035 - val_accuracy: 0.4583
Epoch 11/50
15/15 [==============================] - 9s 639ms/step - loss: 1.0204 -
accuracy: 0.5002 - val_loss: 0.9949 - val_accuracy: 0.4750
Epoch 12/50
15/15 [==============================] - 10s 669ms/step - loss: 0.9873 -
accuracy: 0.4685 - val_loss: 1.0216 - val_accuracy: 0.4333
Epoch 13/50
15/15 [==============================] - 9s 634ms/step - loss: 1.0197 -
accuracy: 0.3335 - val_loss: 1.0399 - val_accuracy: 0.5667
Epoch 14/50
15/15 [==============================] - 10s 640ms/step - loss: 1.0551 -
accuracy: 0.3957 - val_loss: 0.9945 - val_accuracy: 0.4833
Epoch 15/50
15/15 [==============================] - 10s 694ms/step - loss: 1.0044 -
accuracy: 0.4820 - val_loss: 0.9825 - val_accuracy: 0.5167
Epoch 16/50
15/15 [==============================] - 10s 651ms/step - loss: 1.0584 -
accuracy: 0.4248 - val_loss: 1.0339 - val_accuracy: 0.5500
Epoch 17/50
15/15 [==============================] - 9s 635ms/step - loss: 1.0109 -
accuracy: 0.4612 - val_loss: 0.9849 - val_accuracy: 0.5417
Epoch 18/50
15/15 [==============================] - 10s 679ms/step - loss: 1.0051 -
accuracy: 0.5294 - val_loss: 0.9671 - val_accuracy: 0.4833
Epoch 19/50
15/15 [==============================] - 10s 649ms/step - loss: 0.9986 -

5
accuracy: 0.4820 - val_loss: 1.0205 - val_accuracy: 0.5667
Epoch 20/50
15/15 [==============================] - 10s 663ms/step - loss: 1.0622 -
accuracy: 0.4248 - val_loss: 0.9629 - val_accuracy: 0.5417
Epoch 21/50
15/15 [==============================] - 9s 641ms/step - loss: 1.0076 -
accuracy: 0.4397 - val_loss: 0.9544 - val_accuracy: 0.4750
Epoch 22/50
15/15 [==============================] - 10s 666ms/step - loss: 1.0086 -
accuracy: 0.5014 - val_loss: 0.9682 - val_accuracy: 0.4667
Epoch 23/50
15/15 [==============================] - 9s 630ms/step - loss: 0.9784 -
accuracy: 0.5020 - val_loss: 0.9433 - val_accuracy: 0.5167
Epoch 24/50
15/15 [==============================] - 10s 649ms/step - loss: 1.0181 -
accuracy: 0.4546 - val_loss: 0.9826 - val_accuracy: 0.5083
Epoch 25/50
15/15 [==============================] - 9s 643ms/step - loss: 1.0044 -
accuracy: 0.4791 - val_loss: 0.9998 - val_accuracy: 0.5750
Epoch 26/50
15/15 [==============================] - 10s 642ms/step - loss: 1.0451 -
accuracy: 0.4416 - val_loss: 0.9144 - val_accuracy: 0.5667
Epoch 27/50
15/15 [==============================] - 9s 641ms/step - loss: 0.9963 -
accuracy: 0.5264 - val_loss: 0.9123 - val_accuracy: 0.5417
Epoch 28/50
15/15 [==============================] - 10s 659ms/step - loss: 0.9689 -
accuracy: 0.5456 - val_loss: 0.9240 - val_accuracy: 0.5333
Epoch 29/50
15/15 [==============================] - 10s 650ms/step - loss: 1.0336 -
accuracy: 0.4201 - val_loss: 0.9357 - val_accuracy: 0.6000
Epoch 30/50
15/15 [==============================] - 9s 641ms/step - loss: 0.9533 -
accuracy: 0.5617 - val_loss: 0.8898 - val_accuracy: 0.5750
Epoch 31/50
15/15 [==============================] - 10s 643ms/step - loss: 0.9382 -
accuracy: 0.5345 - val_loss: 0.8749 - val_accuracy: 0.5667
Epoch 32/50
15/15 [==============================] - 10s 652ms/step - loss: 0.9554 -
accuracy: 0.5353 - val_loss: 0.8602 - val_accuracy: 0.4917
Epoch 33/50
15/15 [==============================] - 10s 660ms/step - loss: 0.9175 -
accuracy: 0.5476 - val_loss: 0.8537 - val_accuracy: 0.6333
Epoch 34/50
15/15 [==============================] - 9s 633ms/step - loss: 0.9401 -
accuracy: 0.5971 - val_loss: 0.8074 - val_accuracy: 0.5583
Epoch 35/50
15/15 [==============================] - 10s 653ms/step - loss: 0.8931 -

6
accuracy: 0.5488 - val_loss: 0.8308 - val_accuracy: 0.6500
Epoch 36/50
15/15 [==============================] - 10s 652ms/step - loss: 0.9257 -
accuracy: 0.4885 - val_loss: 0.8838 - val_accuracy: 0.6167
Epoch 37/50
15/15 [==============================] - 10s 646ms/step - loss: 0.9359 -
accuracy: 0.5518 - val_loss: 0.8156 - val_accuracy: 0.5167
Epoch 38/50
15/15 [==============================] - 10s 642ms/step - loss: 0.9125 -
accuracy: 0.5101 - val_loss: 0.8150 - val_accuracy: 0.6167
Epoch 39/50
15/15 [==============================] - 10s 658ms/step - loss: 0.7747 -
accuracy: 0.6393 - val_loss: 1.0976 - val_accuracy: 0.4250
Epoch 40/50
15/15 [==============================] - 9s 627ms/step - loss: 0.9141 -
accuracy: 0.5028 - val_loss: 0.8036 - val_accuracy: 0.6750
Epoch 41/50
15/15 [==============================] - 10s 655ms/step - loss: 0.8790 -
accuracy: 0.5407 - val_loss: 0.6348 - val_accuracy: 0.6750
Epoch 42/50
15/15 [==============================] - 10s 653ms/step - loss: 1.0064 -
accuracy: 0.4193 - val_loss: 0.8503 - val_accuracy: 0.6250
Epoch 43/50
15/15 [==============================] - 10s 632ms/step - loss: 0.7850 -
accuracy: 0.6380 - val_loss: 0.7612 - val_accuracy: 0.6417
Epoch 44/50
15/15 [==============================] - 10s 658ms/step - loss: 0.8648 -
accuracy: 0.5755 - val_loss: 0.7752 - val_accuracy: 0.6667
Epoch 45/50
15/15 [==============================] - 10s 655ms/step - loss: 0.8568 -
accuracy: 0.5595 - val_loss: 0.7430 - val_accuracy: 0.6083
Epoch 46/50
15/15 [==============================] - 9s 642ms/step - loss: 0.8237 -
accuracy: 0.5996 - val_loss: 0.8745 - val_accuracy: 0.5667
Epoch 47/50
15/15 [==============================] - 10s 663ms/step - loss: 0.9220 -
accuracy: 0.5834 - val_loss: 0.6625 - val_accuracy: 0.6667
Epoch 48/50
15/15 [==============================] - 10s 644ms/step - loss: 0.9779 -
accuracy: 0.5795 - val_loss: 0.6604 - val_accuracy: 0.6417
Epoch 49/50
15/15 [==============================] - 10s 640ms/step - loss: 0.7130 -
accuracy: 0.6859 - val_loss: 0.6824 - val_accuracy: 0.6750
Epoch 50/50
15/15 [==============================] - 9s 638ms/step - loss: 0.7852 -
accuracy: 0.6131 - val_loss: 0.7123 - val_accuracy: 0.6833

7
[12]: # Plot training and validation graphs
acc = history.history['accuracy']
val_accuracy = history.history['val_accuracy']

loss = history.history['loss']
val_loss = history.history['val_loss']

epochs_range = range(EPOCHS)

plt.figure(figsize=(12,12))
plt.subplot(1,2,1)
plt.plot(epochs_range,acc,label='Training Accuracy')
plt.plot(epochs_range,val_accuracy,label='Validation Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')

plt.subplot(1,2,2)
plt.plot(epochs_range,loss,label='Training Loss')
plt.plot(epochs_range,val_loss,label='Validation Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()

8
[13]: from tensorflow.keras.models import load_model
from tensorflow.keras.preprocessing import image