ForgeryDetection.py

from sklearn.cluster import DBSCAN
import numpy as np
import cv2


class Detect(object):
    def __init__(self, input):
        self.image = cv2.imread(input)

    def siftDetector(self):
        sift = cv2.SIFT_create()
        # sift = cv2.xfeatures2d.SIFT_create()
        gray = cv2.cvtColor(self.image, cv2.COLOR_BGR2GRAY)
        self.key_points, self.descriptors = sift.detectAndCompute(gray, None)
        return self.key_points, self.descriptors

    def showSiftFeatures(self):
        gray_image = cv2.cvtColor(self.image, cv2.COLOR_BGR2GRAY)
        sift_image = cv2.drawKeypoints(
            self.image, self.key_points, self.image.copy())
        return sift_image

    def locateForgery(self, eps=40, min_sample=2):
        clusters = DBSCAN(eps=eps, min_samples=min_sample).fit(
            self.descriptors)
        size = np.unique(clusters.labels_).shape[0]-1
        forgery = self.image.copy()
        if (size == 0) and (np.unique(clusters.labels_)[0] == -1):
            print('No Forgery Found!!')
            return None
        if size == 0:
            size = 1
        cluster_list = [[] for i in range(size)]
        for idx in range(len(self.key_points)):
            if clusters.labels_[idx] != -1:
                cluster_list[clusters.labels_[idx]].append(
                    (int(self.key_points[idx].pt[0]), int(self.key_points[idx].pt[1])))
        for points in cluster_list:
            if len(points) > 1:
                for idx1 in range(1, len(points)):
                    # Green color in BGR
                    cv2.line(forgery, points[0], points[idx1], (0, 255, 0), 5)
                    # cv2.line(forgery, points[0], points[idx1], (255, 0, 0), 5)
        return forgery