Density-Based Clustering with Kernel Diffusion

TL;DR

This paper introduces a new adaptive kernel diffusion density function for density-based clustering, improving local feature capture and computational efficiency, leading to superior clustering performance on benchmark and face datasets.

Contribution

The paper proposes a novel kernel diffusion density function and an efficient surrogate, enhancing density-based clustering's adaptability and scalability.

Findings

Significant improvement over classic density-based algorithms

Outperforms state-of-the-art face clustering methods

Efficient linear-time computation

Abstract

Finding a suitable density function is essential for density-based clustering algorithms such as DBSCAN and DPC. A naive density corresponding to the indicator function of a unit $d$-dimensional Euclidean ball is commonly used in these algorithms. Such density suffers from capturing local features in complex datasets. To tackle this issue, we propose a new kernel diffusion density function, which is adaptive to data of varying local distributional characteristics and smoothness. Furthermore, we develop a surrogate that can be efficiently computed in linear time and space and prove that it is asymptotically equivalent to the kernel diffusion density function. Extensive empirical experiments on benchmark and large-scale face image datasets show that the proposed approach not only achieves a significant improvement over classic density-based clustering algorithms but also outperforms the state-of-the-art face clustering methods by a large margin.