Text this: Single-Character-Based Embedding Feature Aggregation Using Cross-Attention for Scene Text Super-Resolution