Self-Correcting Text-to-Video Generation with Misalignment Detection and Localized Refinement