Why image super-resolution do not use colour information

Read many papers on SR, however, I found almost all of them only use intensity for reconstruction and evaluation. I am wondering why do not use some colour information?