Search papers, labs, and topics across Lattice.
The paper introduces Predictive Spectral Calibration (PSC), a source-free test-time adaptation (TTA) method for image regression that addresses the limitations of adapting classification-specific TTA techniques to continuous regression targets. PSC extends subspace alignment to block spectral matching, aligning target features within the source predictive support and calibrating residual spectral slack in the orthogonal complement. Experiments on image regression benchmarks demonstrate that PSC consistently outperforms strong baselines, especially under severe distribution shifts, while remaining simple to implement and model-agnostic.
Source-free test-time adaptation for image regression gets a boost with Predictive Spectral Calibration, which aligns target features within the source predictive support and calibrates residual spectral slack, leading to significant performance gains under distribution shifts.
Test-time adaptation (TTA) for image regression has received far less attention than its classification counterpart. Methods designed for classification often depend on classification-specific objectives and decision boundaries, making them difficult to transfer directly to continuous regression targets. Recent progress revisits regression TTA through subspace alignment, showing that simple source-guided alignment can be both practical and effective. Building on this line of work, we propose Predictive Spectral Calibration (PSC), a source-free framework that extends subspace alignment to block spectral matching. Instead of relying on a fixed support subspace alone, PSC jointly aligns target features within the source predictive support and calibrates residual spectral slack in the orthogonal complement. PSC remains simple to implement, model-agnostic, and compatible with off-the-shelf pretrained regressors. Experiments on multiple image regression benchmarks show consistent improvements over strong baselines, with particularly clear gains under severe distribution shifts.