Publication

Abstract : The world is currently under the grip of the COVID-19 pandemic. The only possible escape from this pandemic is wearing a mask. Mask checks are being done in most public places. This project brings out a good technique for recognizing a mask on a face. Principal component analysis (PCA) is a dimensionality reduction method which is used in image and signal processing. It uses only the principal components of a dataset and ignores rest of the components. In this paper, three different pre-processing methods (modular, wavelet and a combination of both) have been performed on an image dataset. The resultant data has been processed through PCA. Through comparison of the processed data vectors, the similarity between images has been established. Also, a comparison between different PCA techniques is developed through recognition of distorted faces. Initially, a comparison between different PCA techniques is developed through recognition of distorted faces. Then, a comparison is done between the Modular PCA, Wavelet PCA and Modular-Wavelet PCA techniques by observing the performance for mask recognition. Finally, a conclusion is drawn as to which method is the best in terms of mask recognition.