Δημοσιεύσεις

Recently, several deep learning (DL) approaches have been introduced aiming to cope with image classification tasks. Although these models are effective, they are considered “black-boxes”, as they do not provide explanations or justifications for their results. To deal with the challenge of interpretable image classification, a novel framework, named Interpretable Intuitionistic Fuzzy Cognitive Maps (Ι2FCM), is introduced. Intuitionistic FCMs (iFCMs) have been proposed as an extension of FCMs with the aim of offering a natural mechanism for evaluating the quality of their output through the estimation of hesitancy, a concept resembling human hesitation in decision making. In the context of image classification, hesitancy is considered as a degree of unconfidence with which an image is categorized to a class. To the best of our knowledge this is the first time iFCMs are applied for image classification. Further novel contributions of the introduced framework include the following: a) a feature extraction process focusing on the most informative image regions; b) a learning algorithm for automatic data-driven determination of the intuitionistic fuzzy interconnections of the iFCM, thereby reducing human involvement in defining the graph structure; c) an inherently interpretable classification approach based on image contents, providing understandable explanations of its predictions, using linguistic terms. Furthermore, the proposed Ι2FCM framework can be applied to DL models, including Convolutional Neural Network (CNN), rendering them interpretable. The effectiveness of Ι2FCM was evaluated on public datasets, achieving 92 % accuracy on Caltech-101 and 75 % on Caltech-256. Overall, Ι2FCM delivers competitive classification performance, while offering interpretable inferences.