An improved KNN algorithm-based error identification for Local Binary Pattern Histogram of human emotions

Abstract

This research article presents the real-time implementation of an improved KNN algorithm for identifying error pixels in the Local Binary Pattern Histogram Image used for multi-face emotion recognition. The proposed method focuses on the precise identification of error pixels in the image, while the LBPH recognizes human emotions by implementing it in an FPGA device. Specifically, facial emotion recognition algorithms such as Convolutional Neural Networks (CNNs), Support Vector Machines (SVMs), and Haar cascades are used. However, the LBPH algorithm is simple, easy to implement, and highly efficient. With the inclusion of Artificial Intelligence, the LBPH can be freed from error pixels. This work utilizes a novel improved KNN algorithm to identify the error pixels added to the image. The validity of the proposed method requires error induction at random pixels of the image and verification for recognition by the improved KNN algorithm. The use of an FPGA for the proposed methods proves to be advantageous, with high throughput and low latency.