CP-mtML: Coupled Projection multi-task Metric Learning for Large Scale Face Retrieval

TL;DR

The paper introduces CP-mtML, a scalable multi-task metric learning method for large-scale face retrieval that effectively handles heterogeneous datasets and high-dimensional features, improving accuracy over existing approaches.

Contribution

It presents a novel scalable multi-task metric learning approach that works with high-dimensional features and heterogeneous datasets for large-scale face retrieval.

Findings

Demonstrates improved retrieval performance on identity and age tasks.

Shows scalability to datasets with over a million distractors.

Validates effectiveness with LBP and CNN features.

Abstract

We propose a novel Coupled Projection multi-task Metric Learning (CP-mtML) method for large scale face retrieval. In contrast to previous works which were limited to low dimensional features and small datasets, the proposed method scales to large datasets with high dimensional face descriptors. It utilises pairwise (dis-)similarity constraints as supervision and hence does not require exhaustive class annotation for every training image. While, traditionally, multi-task learning methods have been validated on same dataset but different tasks, we work on the more challenging setting with heterogeneous datasets and different tasks. We show empirical validation on multiple face image datasets of different facial traits, e.g. identity, age and expression. We use classic Local Binary Pattern (LBP) descriptors along with the recent Deep Convolutional Neural Network (CNN) features. The experiments clearly demonstrate the scalability and improved performance of the proposed method on the tasks of identity and age based face image retrieval compared to competitive existing methods, on the standard datasets and with the presence of a million distractor face images.