Color Extraction Module for Unsupervised Image Classification Representation
Alexandru-Toma Andrei, Ovidiu Grigore · 2023
Even though working with unlabeled data has the main advantages of lower preprocessing time and resources, the drawback is that the algorithms implemented have no connection between the output and ground truth, therefore is extremely hard to represent the results. This paper proposes a deployable stand-alone module for extracting the real color of the identified classes for a more intuitive representation of the final results.