<p>Fuzzy c-means (FCM) clustering plays an important role in discovering clustering structures of data. However, FCM is sensitive to noise and outliers, leading to poor clustering results. To improve its robustness, a series of derivative robust algorithms based on FCM were proposed. However, these methods perform poorly on datasets with imbalanced clusters, as FCM tends to balance cluster sizes. Effective methods addressing both robustness and size imbalance in FCM are lacking. Aiming at these two problems, this paper proposes a robust fuzzy local information C-means clustering algorithm with size-insensitive property (siRFLICM). Firstly, marginal discriminative information between clusters is incorporated via a local fuzzy clustering model. Then, by integrating this model into FLICM, we propose a superior robust fuzzy clustering algorithm that enhances intra-cluster cohesion and inter-cluster separability by considering both global and local data structures. Finally, median membership filtering defines the “imbalanced” boundary, balancing noise removal and imbalanced data handling. Experiments on artificial datasets and real images demonstrate that the proposed method effectively addresses size imbalance and exhibits strong robustness compared with state-of-the-art FCM methods.</p>