Technical Report v1.11: In a 655.9-Million-Block Database, SSIM's Structure Score Hits a Guaranteed 1.0 in 19.6 Million Cases — And Hides Hallucinated Detail

This report proves a closed-form identity in the Structural Similarity Index (SSIM): when a ground-truth image block is perfectly flat (σgt=0), the SSIM structure term S is mathematically forced to exactly 1, regardless of what the reconstructed block contains — not an approximation, an exact algebraic consequence of the covariance formula. The report validates this identity against every σgt=0 block in the project's full evaluation database (163, 971, 072 blocks per model, 655, 884, 288 across all four AI upscaling models under both noTTA and TTA combined): 19, 688, 824 such blocks (about 3. 0% of the full database), finding zero exceptions across all eight model×mode combinations — verified as a direct block-level check, not an aggregate count comparison. It then checks whether reconstructions actually place non-trivial, hallucinated content into blocks where this blind spot applies. They do, to substantially different degrees across models: Upscayl High-Fidelity's flat-ground-truth population shows hallucinated detail (reconstructed-block sharpness) with a maximum nearly 23× SwinIR's maximum, a difference S is mathematically incapable of registering. All four affected models — Upscayl Standard, SwinIR, Upscayl High-Fidelity, and Upscayl Digital-Art — were verified via live SQL execution against the same databases underlying Technical Report v1. 10 (SHA-256 hashes confirmed identical, since no regeneration occurred since then), screen-recorded in a single continuous session covering all four models (each recording begins by showing Windows system information to establish the recording's authenticity before the SQL execution itself). The complete verification SQL is included to enable independent reproduction. Dataset and Mandatory Citation: - Source: NIH ChestX-ray8 (Hospital-scale chest x-ray database) - Citation: Wang, X. , Peng, Y. , Lu, L. , Lu, Z. , Bagheri, M. , & Summers, R. M. (2017). "ChestX-ray8: Hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thoracic diseases. " Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 3462–3471. - Download: https: //nihcc. app. box. com/v/ChestXray-NIHCC Contact: s. shiny. n. works@gmail. com