A Unified Approach to Kinship Verification
This work addresses kinship verification, a domain-specific problem in computer vision, with incremental improvements in handling small datasets and imbalance.
The paper tackled kinship verification by proposing a unified multi-task learning approach that jointly learns all kinship classes to better utilize small training sets, and it outperformed state-of-the-art results on Families In the Wild, FG2018, and FG2020 datasets.
In this work, we propose a deep learning-based approach for kin verification using a unified multi-task learning scheme where all kinship classes are jointly learned. This allows us to better utilize small training sets that are typical of kin verification. We introduce a novel approach for fusing the embeddings of kin images, to avoid overfitting, which is a common issue in training such networks. An adaptive sampling scheme is derived for the training set images to resolve the inherent imbalance in kin verification datasets. A thorough ablation study exemplifies the effectivity of our approach, which is experimentally shown to outperform contemporary state-of-the-art kin verification results when applied to the Families In the Wild, FG2018, and FG2020 datasets.