A Source-Free Approach for Domain Adaptation via Multiview Image Transformation and Latent Space Consistency

2026-01-28

AI Summary

This work addresses the performance degradation in cross-domain transfer under source-free settings, where distributional shifts hinder adaptation. To tackle this challenge, the authors propose a self-supervised domain adaptation method that relies solely on target-domain data. By integrating multi-view image augmentation with latent-space consistency constraints, the approach jointly optimizes classification loss and multi-view feature consistency to learn domain-invariant representations in an end-to-end manner. Notably, this is the first method in the source-free paradigm to combine multi-view augmentation with latent consistency without requiring source data, adversarial training, or pseudo-labeling. Using a ConvNeXt encoder, the method achieves state-of-the-art average accuracies of 90.72%, 84.0%, and 97.12% on Office-31, Office-Home, and Office-Caltech, respectively, significantly outperforming existing source-free domain adaptation approaches.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Multi-modal Vision
• Application Domains: Humanities & Computational Social Science

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Security and Privacy: Data transparency and provenance

Abstract

Domain adaptation (DA) addresses the challenge of transferring knowledge from a source domain to a target domain where image data distributions may differ. Existing DA methods often require access to source domain data, adversarial training, or complex pseudo-labeling techniques, which are computationally expensive. To address these challenges, this paper introduces a novel source-free domain adaptation method. It is the first approach to use multiview augmentation and latent space consistency techniques to learn domain-invariant features directly from the target domain. Our method eliminates the need for source-target alignment or pseudo-label refinement by learning transferable representations solely from the target domain by enforcing consistency between multiple augmented views in the latent space. Additionally, the method ensures consistency in the learned features by generating multiple augmented views of target domain data and minimizing the distance between their feature representations in the latent space. We also introduce a ConvNeXt-based encoder and design a loss function that combines classification and consistency objectives to drive effective adaptation directly from the target domain. The proposed model achieves an average classification accuracy of 90. 72\%, 84\%, and 97. 12\% in Office-31, Office-Home and Office-Caltech datasets, respectively. Further evaluations confirm that our study improves existing methods by an average classification accuracy increment of +1.23\%, +7.26\%, and +1.77\% on the respective datasets.

Problem

• domain adaptation
• source-free
• latent space consistency
• multiview augmentation
• feature representation

Innovation

• Source-Free Domain Adaptation
• Multiview Augmentation
• Latent Space Consistency
• Domain-Invariant Features
• ConvNeXt Encoder