SSDA: Bridging Spectral and Structural Gaps via Dual Adaptation for Vision-Based Time Series Forecasting

SSDA identifies and addresses spectral and structural gaps between rendered time series images and natural images, proposing dual adaptation mechanisms to improve vision model transfer for time series forecasting tasks.