Diagram of Pix2PixCC model and adversarial training for WL-to-CaIIK translation. The generator, G, converts a WL image to the CaIIK modality, the discriminator, D, distinguishes original WL-CaIIK pairs from forged pairs. The inspector(s) contributes to the update of the weights of G by computing correlation coefficients between the original CaIIK image and forged CaIIK image, at multiple scales.