Extended Data Fig. 8: Performance evaluation of spectral signature prediction and unmixing residuals by EndmemberNet using manual processing as ground truth. | Nature Photonics

Extended Data Fig. 8: Performance evaluation of spectral signature prediction and unmixing residuals by EndmemberNet using manual processing as ground truth.

From: Excitation-encoded single-emission shortwave infrared lanthanide fluorophore palette for real-time in vivo multispectral imaging

Extended Data Fig. 8: Performance evaluation of spectral signature prediction and unmixing residuals by EndmemberNet using manual processing as ground truth.

Each row displays, from left to right: Spectral signatures extracted by EndmemberNet compared to those obtained through manual selection; Relative residual maps for 9 examples (another example is presented in Fig. 4) in the test datasets, highlighting the comparison between EndmemberNet results and manual ground truth.

Back to article page