This document presents a Khmer optical character recognition system that uses generic Fourier descriptors for feature extraction and recognition. It discusses pre-processing techniques like binarization and noise removal, segmentation using connected component analysis, and extracting generic Fourier descriptors to represent each character as a feature vector. During recognition, the system compares feature vectors of input characters to those in a database to find the closest match and output the recognized text. The goal is to build a reliable Khmer OCR system that is independent of font and size.