The rapid advancement of AI, Machine Learning and deep learning techniques has resulted in new techniques and variety of tools are used for the manipulation of multimedia. Although initially, the technology was mostly used in legitimate applications like entertainment and education, malicious users have exploited them for unlawful purposes. The growth in deep learning technologies and social media platforms has led to a significant increase in the generation of deepfake content. Deepfake content is widely used for spreading fake information, defamation, and creating and altering vulgar content. Various techniques have been proposed for deepfake detection in multimedia content, utilizing machine learning (ML) and deep learning (DL) approaches. This paper provides the implementation of the deepfake detection in facial images using local binary pattern (LBP) based texture descriptor and histogram of oriented gradient-based shape descriptor (HOG) using different ML classifiers such as K-nearest neighbour (KNN), classification Tree and Support vector machine (SVM). It has been observed that when LBP and HOG features are integrated the accuracy in deepfake detection is significantly improved. The LBP+HOG feature collaboration provides overall improved accuracy of 98.8% for SVM, 91.8% for CT and 98.6% for CT classifier for deepfake dataset.