Effects of Distance Metrics on Color Recognition using Color Histogram and KNN Algorithm
Ishika Saini, Vikas Maheshkar · 2022 IEEE Delhi Section Conference (DELCON) · 2022
Color recognition is crucial in image processing for color-based applications such as object recognition, face recognition, skin color recognition, traffic signal color detection, food color code detection, and so on. The application of classification algorithms, such as the KNN algorithm, facilitates the ability to distinguish between different colors. However, the accuracy of the classification may vary based on the similarity metrics employed. The effect of four different similarity metrics - Euclidean, Chi Square, Chebyshev, and Hamming distance metrics - on the accuracy of color recognition using the RGB Color Histogram for feature extraction and the KNN algorithm for classification among six distinct colors for values of K= 3, 5, and 7 is compared in this paper.