Brain Tumor Detection using Pretrained Model

using Attention Mechanism

import numpy as np
import pandas as pd

base_path =
"/kaggle/input/brain-tumour-classification/BrainTumor_1/Train"
categories = ["glioma", "meningioma", "notumor", "pituitary"]

image_paths = []
labels = []

for category in categories:

category_path = os.path.join(base_path, category)
for image_name in os.listdir(category_path):
image_path = os.path.join(category_path, image_name)
 image_paths.append(image_path)
labels.append(category)

df = pd.DataFrame({
"image_path": image_paths,
"label": labels
})

df.head()

image_path label
0 /kaggle/input/brain-tumour-classification/Brai... glioma
1 /kaggle/input/brain-tumour-classification/Brai... glioma
2 /kaggle/input/brain-tumour-classification/Brai... glioma
3 /kaggle/input/brain-tumour-classification/Brai... glioma
4 /kaggle/input/brain-tumour-classification/Brai... glioma

df.tail()

image_path label
22843 /kaggle/input/brain-tumour-classification/Brai... pituitary
22844 /kaggle/input/brain-tumour-classification/Brai... pituitary
22845 /kaggle/input/brain-tumour-classification/Brai... pituitary
22846 /kaggle/input/brain-tumour-classification/Brai... pituitary
22847 /kaggle/input/brain-tumour-classification/Brai... pituitary

df.shape

(22848, 2)

df.columns

Index(['image_path', 'label'], dtype='object')

df['label'].unique()

array(['glioma', 'meningioma', 'notumor', 'pituitary'], dtype=object)

df['label'].value_counts()

label
notumor 6380
pituitary 5828
meningioma 5356
glioma 5284
Name: count, dtype: int64

import seaborn as sns

import matplotlib.pyplot as plt

plt.figure(figsize=(8, 6))
sns.countplot(data=df, x="label", palette="viridis")
plt.title("Distribution of Labels - Count Plot")
plt.xlabel("Tumor Type")
plt.ylabel("Count")

for p in plt.gca().patches:
plt.gca().annotate(f'{int(p.get_height())}',
(p.get_x() + p.get_width() / 2.,
p.get_height()),
ha='center', va='center', fontsize=11,
color='black', xytext=(0, 5),
textcoords='offset points')

plt.show()

label_counts = df['label'].value_counts()
plt.figure(figsize=(8, 6))
plt.pie(label_counts, labels=label_counts.index, autopct='%1.1f%%',
startangle=140, colors=sns.color_palette("viridis"))
plt.title("Distribution of Labels - Pie Chart")
plt.show()
import cv2

num_images = 5

plt.figure(figsize=(15, 12))

for i, category in enumerate(categories):

category_images = df[df['label'] == category]
['image_path'].iloc[:num_images]

for j, img_path in enumerate(category_images):

img = cv2.imread(img_path)
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

plt.subplot(len(categories), num_images, i * num_images + j +

1)
plt.imshow(img)
plt.axis('off')
plt.title(category)
plt.tight_layout()
plt.show()

from sklearn.preprocessing import LabelEncoder

label_encoder = LabelEncoder()

df['category_encoded'] = label_encoder.fit_transform(df['label'])

df = df[['image_path', 'category_encoded']]

from imblearn.over_sampling import RandomOverSampler

ros = RandomOverSampler(random_state=42)
X_resampled, y_resampled = ros.fit_resample(df[['image_path']],
df['category_encoded'])

df_resampled = pd.DataFrame(X_resampled, columns=['image_path'])

df_resampled['category_encoded'] = y_resampled
print("\nClass distribution after oversampling:")
print(df_resampled['category_encoded'].value_counts())

Class distribution after oversampling:

category_encoded
0 6380
1 6380
2 6380
3 6380
Name: count, dtype: int64

df_resampled

image_path
category_encoded
0 /kaggle/input/brain-tumour-classification/Brai...
0
1 /kaggle/input/brain-tumour-classification/Brai...
0
2 /kaggle/input/brain-tumour-classification/Brai...
0
3 /kaggle/input/brain-tumour-classification/Brai...
0
4 /kaggle/input/brain-tumour-classification/Brai...
0
... ...
...
25515 /kaggle/input/brain-tumour-classification/Brai...
3
25516 /kaggle/input/brain-tumour-classification/Brai...
3
25517 /kaggle/input/brain-tumour-classification/Brai...
3
25518 /kaggle/input/brain-tumour-classification/Brai...
3
25519 /kaggle/input/brain-tumour-classification/Brai...
3

[25520 rows x 2 columns]

import time
import shutil
import pathlib
import itertools
from PIL import Image

import cv2
import seaborn as sns
sns.set_style('darkgrid')
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix, classification_report

import tensorflow as tf
from tensorflow import keras
from tensorflow.keras.models import Sequential
from tensorflow.keras.optimizers import Adam, Adamax
from tensorflow.keras.preprocessing.image import ImageDataGenerator
from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten,
Dense, Activation, Dropout, BatchNormalization
from tensorflow.keras import regularizers

import warnings
warnings.filterwarnings("ignore")

print ('check')

check

df_resampled['category_encoded'] =
df_resampled['category_encoded'].astype(str)

train_df_new, temp_df_new = train_test_split(

df_resampled,
train_size=0.8,
shuffle=True,
random_state=42,
stratify=df_resampled['category_encoded']
)

valid_df_new, test_df_new = train_test_split(

temp_df_new,
test_size=0.5,
shuffle=True,
random_state=42,
stratify=temp_df_new['category_encoded']
)

from tensorflow.keras.preprocessing.image import ImageDataGenerator

batch_size = 16
img_size = (224, 224)
channels = 3
img_shape = (img_size[0], img_size[1], channels)

tr_gen = ImageDataGenerator(rescale=1./255)
ts_gen = ImageDataGenerator(rescale=1./255)

train_gen_new = tr_gen.flow_from_dataframe(
train_df_new,
 x_col='image_path',
y_col='category_encoded',
target_size=img_size,
class_mode='sparse',
color_mode='rgb',
shuffle=True,
batch_size=batch_size
)

valid_gen_new = ts_gen.flow_from_dataframe(
valid_df_new,
x_col='image_path',
y_col='category_encoded',
target_size=img_size,
class_mode='sparse',
color_mode='rgb',
shuffle=True,
batch_size=batch_size
)

test_gen_new = ts_gen.flow_from_dataframe(
test_df_new,
x_col='image_path',
y_col='category_encoded',
target_size=img_size,
class_mode='sparse',
color_mode='rgb',
shuffle=False,
batch_size=batch_size
)

Found 20416 validated image filenames belonging to 4 classes.

Found 2552 validated image filenames belonging to 4 classes.
Found 2552 validated image filenames belonging to 4 classes.

import tensorflow as tf
from tensorflow.keras import layers, models
from tensorflow.keras.callbacks import EarlyStopping, ModelCheckpoint

physical_devices = tf.config.list_physical_devices('GPU')
if physical_devices:
print("Using GPU")
else:
print("Using CPU")

Using GPU

early_stopping = EarlyStopping(monitor='val_loss', patience=5,

restore_best_weights=True)
from tensorflow.keras.applications import VGG16
from tensorflow.keras.models import Model
from tensorflow.keras.layers import (GlobalAveragePooling2D, Dense,
Dropout, BatchNormalization,
GaussianNoise, Input,
MultiHeadAttention, Reshape)
from tensorflow.keras.optimizers import Adam

def create_vgg16_model(input_shape):

inputs = Input(shape=input_shape)

base_model = VGG16(weights='imagenet', input_tensor=inputs,

include_top=False)

for layer in base_model.layers:

layer.trainable = False

x = base_model.output

height, width, channels = 7, 7, 512

x = Reshape((height * width, channels))(x)

attention_output = MultiHeadAttention(num_heads=8,
key_dim=channels)(x, x)
attention_output = Reshape((height, width, channels))
(attention_output)

x = GaussianNoise(0.25)(attention_output)

x = GlobalAveragePooling2D()(x)

x = Dense(512, activation='relu')(x)
x = BatchNormalization()(x)
x = GaussianNoise(0.25)(x)
x = Dropout(0.25)(x)

outputs = Dense(4, activation='softmax')(x)

model = Model(inputs=inputs, outputs=outputs)

return model

input_shape = (224, 224, 3)

cnn_model = create_vgg16_model(input_shape)

cnn_model.compile(optimizer=Adam(learning_rate=0.0001),
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
history = cnn_model.fit(
train_gen_new,
validation_data=valid_gen_new,
epochs=5,
callbacks=[early_stopping],
verbose=1
)

Epoch 1/5

WARNING: All log messages before absl::InitializeLog() is called are

written to STDERR
I0000 00:00:1730616392.742667 299 service.cc:145] XLA service
0x7a917c0018a0 initialized for platform CUDA (this does not guarantee
that XLA will be used). Devices:
I0000 00:00:1730616392.742725 299 service.cc:153] StreamExecutor
device (0): Tesla T4, Compute Capability 7.5
I0000 00:00:1730616392.742730 299 service.cc:153] StreamExecutor
device (1): Tesla T4, Compute Capability 7.5
WARNING: All log messages before absl::InitializeLog() is called are
written to STDERR
I0000 00:00:1730616404.835907 338 asm_compiler.cc:369] ptxas
warning : Registers are spilled to local memory in function
'triton_gemm_dot_1', 1284 bytes spill stores, 1272 bytes spill loads

1/1276 ━━━━━━━━━━━━━━━━━━━━ 9:55:19 28s/step - accuracy: 0.2500 -

loss: 1.9726

I0000 00:00:1730616416.859234 299 device_compiler.h:188] Compiled

cluster using XLA! This line is logged at most once for the lifetime
of the process.

1276/1276 ━━━━━━━━━━━━━━━━━━━━ 0s 78ms/step - accuracy: 0.7626 - loss:

0.6298

I0000 00:00:1730616537.083262 399 asm_compiler.cc:369] ptxas

warning : Registers are spilled to local memory in function
'triton_gemm_dot_6', 1280 bytes spill stores, 1268 bytes spill loads

1276/1276 ━━━━━━━━━━━━━━━━━━━━ 154s 99ms/step - accuracy: 0.7627 -

loss: 0.6296 - val_accuracy: 0.6759 - val_loss: 1.0673
Epoch 2/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 116s 90ms/step - accuracy: 0.8798 -
loss: 0.3314 - val_accuracy: 0.8973 - val_loss: 0.2846
Epoch 3/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 116s 90ms/step - accuracy: 0.9128 -
loss: 0.2491 - val_accuracy: 0.8997 - val_loss: 0.2999
Epoch 4/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 116s 90ms/step - accuracy: 0.9180 -
loss: 0.2193 - val_accuracy: 0.7022 - val_loss: 1.3966
Epoch 5/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 116s 90ms/step - accuracy: 0.9317 -
loss: 0.1875 - val_accuracy: 0.5247 - val_loss: 2.6919

plt.plot(history.history['accuracy'])
plt.plot(history.history['val_accuracy'])
plt.title('Model accuracy')
plt.ylabel('Accuracy')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper left')
plt.show()

plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('Model loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper left')
plt.show()
test_labels = test_gen_new.classes
predictions = cnn_model.predict(test_gen_new)
predicted_classes = np.argmax(predictions, axis=1)

160/160 ━━━━━━━━━━━━━━━━━━━━ 13s 76ms/step

report = classification_report(test_labels, predicted_classes,

target_names=list(test_gen_new.class_indices.keys()))
print(report)

precision recall f1-score support

0 0.88 0.93 0.90 638

1 0.82 0.87 0.85 638
2 0.95 0.98 0.97 638
3 0.99 0.85 0.91 638

accuracy 0.91 2552

macro avg 0.91 0.91 0.91 2552
weighted avg 0.91 0.91 0.91 2552

conf_matrix = confusion_matrix(test_labels, predicted_classes)

plt.figure(figsize=(10, 8))
sns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues',
xticklabels=list(test_gen_new.class_indices.keys()),
yticklabels=list(test_gen_new.class_indices.keys()))
plt.title('Confusion Matrix')
plt.xlabel('Predicted Label')
plt.ylabel('True Label')
plt.show()

from tensorflow.keras.applications import VGG19

from tensorflow.keras.models import Model
from tensorflow.keras.layers import GlobalAveragePooling2D, Dense,
Dropout, BatchNormalization, GaussianNoise, Input, MultiHeadAttention,
Reshape
from tensorflow.keras.optimizers import Adam
import tensorflow as tf

def create_vgg19_model(input_shape):

inputs = Input(shape=input_shape)

base_model = VGG19(weights='imagenet', input_tensor=inputs,

include_top=False)

for layer in base_model.layers:

layer.trainable = False

x = base_model.output

height, width, channels = 7, 7, 512

x = Reshape((height * width, channels))(x)

attention_output = MultiHeadAttention(num_heads=8,
key_dim=channels)(x, x)

attention_output = Reshape((height, width, channels))

(attention_output)

x = GaussianNoise(0.25)(attention_output)

x = GlobalAveragePooling2D()(x)

x = Dense(512, activation='relu')(x)
x = BatchNormalization()(x)
x = GaussianNoise(0.25)(x)
x = Dropout(0.25)(x)

outputs = Dense(4, activation='softmax')(x)

model = Model(inputs=inputs, outputs=outputs)

return model

input_shape = (224, 224, 3)

cnn_model = create_vgg19_model(input_shape)

cnn_model.compile(optimizer=Adam(learning_rate=0.0001),
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])

Downloading data from https://storage.googleapis.com/tensorflow/keras-

applications/vgg19/vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5
80134624/80134624 ━━━━━━━━━━━━━━━━━━━━ 0s 0us/step
history = cnn_model.fit(
train_gen_new,
validation_data=valid_gen_new,
epochs=5,
callbacks=[early_stopping],
verbose=1
)

Epoch 1/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 149s 110ms/step - accuracy: 0.7411 -
loss: 0.6771 - val_accuracy: 0.7786 - val_loss: 0.6523
Epoch 2/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 138s 108ms/step - accuracy: 0.8712 -
loss: 0.3527 - val_accuracy: 0.8288 - val_loss: 0.5774
Epoch 3/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 138s 108ms/step - accuracy: 0.9022 -
loss: 0.2665 - val_accuracy: 0.8546 - val_loss: 0.3942
Epoch 4/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 138s 108ms/step - accuracy: 0.9215 -
loss: 0.2168 - val_accuracy: 0.8064 - val_loss: 0.5658
Epoch 5/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 138s 108ms/step - accuracy: 0.9319 -
loss: 0.1890 - val_accuracy: 0.9150 - val_loss: 0.2445

plt.plot(history.history['accuracy'])
plt.plot(history.history['val_accuracy'])
plt.title('Model accuracy')
plt.ylabel('Accuracy')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper left')
plt.show()

plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('Model loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper left')
plt.show()
test_labels = test_gen_new.classes
predictions = cnn_model.predict(test_gen_new)
predicted_classes = np.argmax(predictions, axis=1)

160/160 ━━━━━━━━━━━━━━━━━━━━ 15s 91ms/step

report = classification_report(test_labels, predicted_classes,

target_names=list(test_gen_new.class_indices.keys()))
print(report)

precision recall f1-score support

0 0.98 0.81 0.89 638

1 0.83 0.92 0.87 638
2 0.97 0.97 0.97 638
3 0.92 0.98 0.95 638

accuracy 0.92 2552

macro avg 0.92 0.92 0.92 2552
weighted avg 0.92 0.92 0.92 2552

conf_matrix = confusion_matrix(test_labels, predicted_classes)

plt.figure(figsize=(10, 8))
sns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues',
xticklabels=list(test_gen_new.class_indices.keys()),
yticklabels=list(test_gen_new.class_indices.keys()))
plt.title('Confusion Matrix')
plt.xlabel('Predicted Label')
plt.ylabel('True Label')
plt.show()

from tensorflow.keras.applications import MobileNet

from tensorflow.keras.models import Model
from tensorflow.keras.layers import GlobalAveragePooling2D, Dense,
Dropout, BatchNormalization, GaussianNoise, Input, MultiHeadAttention,
Reshape
from tensorflow.keras.optimizers import Adam
import tensorflow as tf

def create_mobilenet_model(input_shape):

inputs = Input(shape=input_shape)

base_model = MobileNet(weights='imagenet', input_tensor=inputs,

include_top=False)

for layer in base_model.layers:

layer.trainable = False

x = base_model.output

height, width, channels = 7, 7, 1024

x = Reshape((height * width, channels))(x)

attention_output = MultiHeadAttention(num_heads=8,
key_dim=channels)(x, x)

attention_output = Reshape((height, width, channels))

(attention_output)

x = GaussianNoise(0.25)(attention_output)

x = GlobalAveragePooling2D()(x)

x = Dense(512, activation='relu')(x)
x = BatchNormalization()(x)
x = GaussianNoise(0.25)(x)
x = Dropout(0.25)(x)

outputs = Dense(4, activation='softmax')(x)

model = Model(inputs=inputs, outputs=outputs)

return model

input_shape = (224, 224, 3)

cnn_model = create_mobilenet_model(input_shape)

cnn_model.compile(optimizer=Adam(learning_rate=0.0001),
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])

Downloading data from https://storage.googleapis.com/tensorflow/keras-

applications/mobilenet/mobilenet_1_0_224_tf_no_top.h5
17225924/17225924 ━━━━━━━━━━━━━━━━━━━━ 0s 0us/step
history = cnn_model.fit(
train_gen_new,
validation_data=valid_gen_new,
epochs=5,
callbacks=[early_stopping],
verbose=1
)

Epoch 1/5

I0000 00:00:1730617968.755983 673 asm_compiler.cc:369] ptxas

warning : Registers are spilled to local memory in function
'triton_gemm_dot_1', 1300 bytes spill stores, 1284 bytes spill loads

1276/1276 ━━━━━━━━━━━━━━━━━━━━ 0s 55ms/step - accuracy: 0.8556 - loss:

0.4049

I0000 00:00:1730618059.939100 726 asm_compiler.cc:369] ptxas

warning : Registers are spilled to local memory in function
'triton_gemm_dot_1', 1300 bytes spill stores, 1284 bytes spill loads

1276/1276 ━━━━━━━━━━━━━━━━━━━━ 109s 68ms/step - accuracy: 0.8557 -

loss: 0.4048 - val_accuracy: 0.9412 - val_loss: 0.1564
Epoch 2/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 77s 60ms/step - accuracy: 0.9400 -
loss: 0.1708 - val_accuracy: 0.8374 - val_loss: 0.5594
Epoch 3/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 76s 59ms/step - accuracy: 0.9632 -
loss: 0.1061 - val_accuracy: 0.9208 - val_loss: 0.2246
Epoch 4/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 76s 59ms/step - accuracy: 0.9756 -
loss: 0.0739 - val_accuracy: 0.9671 - val_loss: 0.1134
Epoch 5/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 76s 60ms/step - accuracy: 0.9779 -
loss: 0.0591 - val_accuracy: 0.9659 - val_loss: 0.1115

plt.plot(history.history['accuracy'])
plt.plot(history.history['val_accuracy'])
plt.title('Model accuracy')
plt.ylabel('Accuracy')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper left')
plt.show()

plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('Model loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper left')
plt.show()
test_labels = test_gen_new.classes
predictions = cnn_model.predict(test_gen_new)
predicted_classes = np.argmax(predictions, axis=1)

160/160 ━━━━━━━━━━━━━━━━━━━━ 10s 52ms/step

report = classification_report(test_labels, predicted_classes,

target_names=list(test_gen_new.class_indices.keys()))
print(report)

precision recall f1-score support

0 0.98 0.97 0.98 638

1 0.93 0.97 0.95 638
2 1.00 0.98 0.99 638
3 0.98 0.98 0.98 638

accuracy 0.97 2552

macro avg 0.97 0.97 0.97 2552
weighted avg 0.97 0.97 0.97 2552

conf_matrix = confusion_matrix(test_labels, predicted_classes)

plt.figure(figsize=(10, 8))
sns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues',
xticklabels=list(test_gen_new.class_indices.keys()),
yticklabels=list(test_gen_new.class_indices.keys()))
plt.title('Confusion Matrix')
plt.xlabel('Predicted Label')
plt.ylabel('True Label')
plt.show()

from tensorflow.keras.applications import Xception

from tensorflow.keras.models import Model
from tensorflow.keras.layers import GlobalAveragePooling2D, Dense,
Dropout, BatchNormalization, GaussianNoise, Input, MultiHeadAttention,
Reshape
from tensorflow.keras.optimizers import Adam
import tensorflow as tf

def create_xception_model(input_shape):

inputs = Input(shape=input_shape)

base_model = Xception(weights='imagenet', input_tensor=inputs,

include_top=False)

for layer in base_model.layers:

layer.trainable = False

x = base_model.output

height, width, channels = 7, 7, 2048

x = Reshape((height * width, channels))(x)

attention_output = MultiHeadAttention(num_heads=8,
key_dim=channels)(x, x)

attention_output = Reshape((height, width, channels))

(attention_output)

x = GaussianNoise(0.25)(attention_output)
x = GlobalAveragePooling2D()(x)
x = Dense(512, activation='relu')(x)
x = BatchNormalization()(x)
x = GaussianNoise(0.25)(x)
x = Dropout(0.25)(x)
outputs = Dense(4, activation='softmax')(x)

model = Model(inputs=inputs, outputs=outputs)

return model

input_shape = (224, 224, 3)

cnn_model = create_xception_model(input_shape)

cnn_model.compile(optimizer=Adam(learning_rate=0.0001),
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])

Downloading data from https://storage.googleapis.com/tensorflow/keras-

applications/xception/
xception_weights_tf_dim_ordering_tf_kernels_notop.h5
83683744/83683744 ━━━━━━━━━━━━━━━━━━━━ 0s 0us/step

history = cnn_model.fit(
train_gen_new,
validation_data=valid_gen_new,
 epochs=5,
callbacks=[early_stopping],
verbose=1
)

Epoch 1/5

I0000 00:00:1730618526.007229 855 asm_compiler.cc:369] ptxas

warning : Registers are spilled to local memory in function
'triton_gemm_dot_1', 1300 bytes spill stores, 1284 bytes spill loads

1276/1276 ━━━━━━━━━━━━━━━━━━━━ 0s 231ms/step - accuracy: 0.7874 -

loss: 0.5876

I0000 00:00:1730618864.402637 915 asm_compiler.cc:369] ptxas

warning : Registers are spilled to local memory in function
'triton_gemm_dot_1', 1300 bytes spill stores, 1284 bytes spill loads

1276/1276 ━━━━━━━━━━━━━━━━━━━━ 362s 260ms/step - accuracy: 0.7874 -

loss: 0.5875 - val_accuracy: 0.9087 - val_loss: 0.2856
Epoch 2/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 317s 248ms/step - accuracy: 0.8945 -
loss: 0.2957 - val_accuracy: 0.9032 - val_loss: 0.2573
Epoch 3/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 315s 247ms/step - accuracy: 0.9206 -
loss: 0.2221 - val_accuracy: 0.8997 - val_loss: 0.2763
Epoch 4/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 315s 247ms/step - accuracy: 0.9414 -
loss: 0.1703 - val_accuracy: 0.9487 - val_loss: 0.1454
Epoch 5/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 315s 247ms/step - accuracy: 0.9532 -
loss: 0.1366 - val_accuracy: 0.9549 - val_loss: 0.1152

plt.plot(history.history['accuracy'])
plt.plot(history.history['val_accuracy'])
plt.title('Model accuracy')
plt.ylabel('Accuracy')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper left')
plt.show()

plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('Model loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper left')
plt.show()
test_labels = test_gen_new.classes
predictions = cnn_model.predict(test_gen_new)
predicted_classes = np.argmax(predictions, axis=1)

160/160 ━━━━━━━━━━━━━━━━━━━━ 25s 139ms/step

report = classification_report(test_labels, predicted_classes,

target_names=list(test_gen_new.class_indices.keys()))
print(report)

precision recall f1-score support

0 0.91 0.94 0.92 638

1 0.87 0.80 0.83 638
2 0.90 0.98 0.94 638
3 0.95 0.90 0.92 638

accuracy 0.91 2552

macro avg 0.91 0.91 0.90 2552
weighted avg 0.91 0.91 0.90 2552

conf_matrix = confusion_matrix(test_labels, predicted_classes)

plt.figure(figsize=(10, 8))
sns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues',
xticklabels=list(test_gen_new.class_indices.keys()),
yticklabels=list(test_gen_new.class_indices.keys()))
plt.title('Confusion Matrix')
plt.xlabel('Predicted Label')
plt.ylabel('True Label')
plt.show()

from tensorflow.keras.applications import InceptionV3

from tensorflow.keras.models import Model
from tensorflow.keras.layers import GlobalAveragePooling2D, Dense,
Dropout, BatchNormalization, GaussianNoise, Input, MultiHeadAttention,
Reshape
from tensorflow.keras.optimizers import Adam
import tensorflow as tf

def create_inception_model(input_shape):

inputs = Input(shape=input_shape)

base_model = InceptionV3(weights='imagenet', input_tensor=inputs,

include_top=False)

for layer in base_model.layers:

layer.trainable = False

x = base_model.output

height, width, channels = 5, 5, 2048

x = Reshape((height * width, channels))(x)

attention_output = MultiHeadAttention(num_heads=8,
key_dim=channels)(x, x)

attention_output = Reshape((height, width, channels))

(attention_output)

x = GaussianNoise(0.25)(attention_output)
x = GlobalAveragePooling2D()(x)
x = Dense(512, activation='relu')(x)
x = BatchNormalization()(x)
x = GaussianNoise(0.25)(x)
x = Dropout(0.25)(x)
outputs = Dense(4, activation='softmax')(x)

model = Model(inputs=inputs, outputs=outputs)

return model

input_shape = (224, 224, 3)

cnn_model = create_inception_model(input_shape)

cnn_model.compile(optimizer=Adam(learning_rate=0.0001),
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])

Downloading data from https://storage.googleapis.com/tensorflow/keras-

applications/inception_v3/
inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5
87910968/87910968 ━━━━━━━━━━━━━━━━━━━━ 0s 0us/step

history = cnn_model.fit(
train_gen_new,
validation_data=valid_gen_new,
epochs=5,
 callbacks=[early_stopping],
verbose=1
)

Epoch 1/5

I0000 00:00:1730620812.574405 1045 asm_compiler.cc:369] ptxas

warning : Registers are spilled to local memory in function
'triton_gemm_dot_1', 1300 bytes spill stores, 1284 bytes spill loads

1276/1276 ━━━━━━━━━━━━━━━━━━━━ 0s 127ms/step - accuracy: 0.7763 -

loss: 0.6269

I0000 00:00:1730621007.694029 1096 asm_compiler.cc:369] ptxas

warning : Registers are spilled to local memory in function
'triton_gemm_dot_1', 1300 bytes spill stores, 1284 bytes spill loads

1276/1276 ━━━━━━━━━━━━━━━━━━━━ 221s 149ms/step - accuracy: 0.7764 -

loss: 0.6267 - val_accuracy: 0.8495 - val_loss: 0.5046
Epoch 2/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 172s 135ms/step - accuracy: 0.8994 -
loss: 0.2807 - val_accuracy: 0.5247 - val_loss: 1.6268
Epoch 3/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 171s 134ms/step - accuracy: 0.9281 -
loss: 0.2047 - val_accuracy: 0.9056 - val_loss: 0.2910
Epoch 4/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 171s 134ms/step - accuracy: 0.9375 -
loss: 0.1739 - val_accuracy: 0.9357 - val_loss: 0.1815
Epoch 5/5
1276/1276 ━━━━━━━━━━━━━━━━━━━━ 172s 134ms/step - accuracy: 0.9478 -
loss: 0.1452 - val_accuracy: 0.7367 - val_loss: 0.9598

plt.plot(history.history['accuracy'])
plt.plot(history.history['val_accuracy'])
plt.title('Model accuracy')
plt.ylabel('Accuracy')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper left')
plt.show()

plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('Model loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['Train', 'Validation'], loc='upper left')
plt.show()
test_labels = test_gen_new.classes
predictions = cnn_model.predict(test_gen_new)
predicted_classes = np.argmax(predictions, axis=1)

160/160 ━━━━━━━━━━━━━━━━━━━━ 20s 94ms/step

report = classification_report(test_labels, predicted_classes,

target_names=list(test_gen_new.class_indices.keys()))
print(report)

precision recall f1-score support

0 0.91 0.90 0.90 638

1 0.93 0.53 0.68 638
2 0.94 0.96 0.95 638
3 0.71 1.00 0.83 638

accuracy 0.85 2552

macro avg 0.87 0.85 0.84 2552
weighted avg 0.87 0.85 0.84 2552

conf_matrix = confusion_matrix(test_labels, predicted_classes)

plt.figure(figsize=(10, 8))
sns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues',
xticklabels=list(test_gen_new.class_indices.keys()),
yticklabels=list(test_gen_new.class_indices.keys()))
plt.title('Confusion Matrix')
plt.xlabel('Predicted Label')
plt.ylabel('True Label')
plt.show()

models = ['VGG16', 'VGG19', 'MobileNet', 'Xception', 'InceptionV3']

precision_scores = [0.91, 0.92, 0.97, 0.91, 0.87]
recall_scores = [0.91, 0.92, 0.97, 0.91, 0.85]
f1_scores = [0.91, 0.92, 0.97, 0.90, 0.84]
accuracy_scores = [0.91, 0.92, 0.97, 0.91, 0.85]

bar_width = 0.2
index = np.arange(len(models))

fig, ax = plt.subplots(figsize=(12, 8))

bars1 = ax.bar(index, precision_scores, bar_width, label='Precision')

bars2 = ax.bar(index + bar_width, recall_scores, bar_width,
label='Recall')
bars3 = ax.bar(index + 2 * bar_width, f1_scores, bar_width, label='F1
Score')
bars4 = ax.bar(index + 3 * bar_width, accuracy_scores, bar_width,
label='Accuracy')

def add_labels(bars):
for bar in bars:
height = bar.get_height()
ax.annotate(f'{height:.2f}', xy=(bar.get_x() + bar.get_width()
/ 2, height),
xytext=(0, 3), textcoords="offset points",
ha='center', va='bottom')

add_labels(bars1)
add_labels(bars2)
add_labels(bars3)
add_labels(bars4)

ax.set_xlabel('Models')
ax.set_ylabel('Scores')
ax.set_title('Comparative Metrics for Different Models')
ax.set_xticks(index + bar_width * 1.5)
ax.set_xticklabels(models)
ax.legend()

plt.tight_layout()
plt.show()
Copyight @ThinkAI - A Machine Learning
Community