This document discusses using the K-nearest neighbors (K-NN) algorithm to detect counterfeit currency. It first introduces the problem of currency counterfeiting and some existing techniques used for detection. It then focuses on analyzing the performance of different machine learning algorithms, finding that K-NN achieved the highest accuracy. The document proceeds to explain how K-NN works and how it can be applied specifically to currency detection by extracting image features and comparing new images to training data. Finally, it acknowledges some disadvantages of K-NN and proposes using convolutional neural networks for larger datasets in future work.