sift_new.py

# Here is an example code for using PCA in SIFT descriptor and then SVM classifier in Python:

# 1. Import necessary libraries:


import cv2
import numpy as np
from sklearn.decomposition import PCA
from sklearn.svm import SVC
from sklearn.metrics import accuracy_score


# 2. Load the dataset and extract SIFT descriptors:


# Load dataset images and labels
train_images = []
train_labels = []
for i in range(1, 11):
    for j in range(1, 6):
        img = cv2.imread(f"dataset/{i}_{j}.jpg", cv2.IMREAD_GRAYSCALE)
        train_images.append(img)
        train_labels.append(i)

# Create SIFT object and extract descriptors from all images
sift = cv2.xfeatures2d.SIFT_create()
descriptors = []
for img in train_images:
    _, des = sift.detectAndCompute(img, None)
    descriptors.append(des)

# 3. Apply PCA to reduce the dimensionality of the descriptors:


# Flatten all descriptors into a single array and apply PCA to reduce dimensionality
descriptors_flat = np.concatenate(descriptors)
pca = PCA(n_components=128)
descriptors_pca = pca.fit_transform(descriptors_flat)


# 4. Train an SVM classifier on the reduced descriptors:

# Train SVM classifier on reduced descriptors using labels as target values
svm = SVC(kernel='linear')
svm.fit(descriptors_pca, train_labels)

# 5. Test the classifier on a test set and calculate accuracy:

# ```python
# Load test set images and labels, extract SIFT descriptors, apply PCA transformation,
# predict labels using SVM classifier, and calculate accuracy score.
test_images = []
test_labels = []
for i in range(1, 11):
    for j in range(6, 11):
        img = cv2.imread(f"dataset/{i}_{j}.jpg", cv2.IMREAD_GRAYSCALE)
        test_images.append(img)
        test_labels.append(i)

descriptors_test = []
for img in test_images:
    _, des = sift.detectAndCompute(img, None)
    descriptors_test.append(des)

descriptors_test_flat = np.concatenate(descriptors_test)
descriptors_test_pca = pca.transform(descriptors_test_flat)

predicted_labels = svm.predict(descriptors_test_pca)
accuracy = accuracy_score(test_labels, predicted_labels)
print(f"Accuracy: {accuracy}")

# Note: This is just an example code and may need to be modified based on your specific requirements.