Meta-Mining Discriminative Samples for Kinship Verification

TL;DR

This paper introduces a novel meta-mining approach for kinship verification that leverages all available data, including unbalanced positive and negative pairs, to improve discriminative sample selection and model training.

Contribution

It proposes a Discriminative Sample Meta-Mining (DSMM) method that automatically learns to re-weight samples from unbalanced data for better kinship verification.

Findings

Outperforms existing methods on KinFaceW-I, KinFaceW-II, TSKinFace, and Cornell Kinship datasets.

Effectively utilizes all possible pairs, including negative samples, for training.

Demonstrates improved accuracy and robustness in kinship verification tasks.

Abstract

Kinship verification aims to find out whether there is a kin relation for a given pair of facial images. Kinship verification databases are born with unbalanced data. For a database with N positive kinship pairs, we naturally obtain N(N-1) negative pairs. How to fully utilize the limited positive pairs and mine discriminative information from sufficient negative samples for kinship verification remains an open issue. To address this problem, we propose a Discriminative Sample Meta-Mining (DSMM) approach in this paper. Unlike existing methods that usually construct a balanced dataset with fixed negative pairs, we propose to utilize all possible pairs and automatically learn discriminative information from data. Specifically, we sample an unbalanced train batch and a balanced meta-train batch for each iteration. Then we learn a meta-miner with the meta-gradient on the balanced meta-train batch. In the end, the samples in the unbalanced train batch are re-weighted by the learned meta-miner to optimize the kinship models. Experimental results on the widely used KinFaceW-I, KinFaceW-II, TSKinFace, and Cornell Kinship datasets demonstrate the effectiveness of the proposed approach.