SCTANet: A Spatial Attention-Guided CNN-Transformer Aggregation Network for Deep Face Image Super-Resolution | Synapse