Neural network-based clustering techniques like self-organizing feature maps (SOM) and adaptive resonance theory (ART) are commonly used. SOM represents high-dimensional data in a low-dimensional form while preserving essence. ART consists of comparison and recognition layers and clusters input patterns in an unsupervised manner. Kernel-based clustering algorithms operate in a potentially infinite dimensional feature space and can form arbitrary cluster shapes, deal with noise/outliers, and estimate the number of clusters without prior topological knowledge.