Image Captioning Using Enhanced Cross-Modal Attention with Multi-Scale Aggregation for Social Hotspot and Public Opinion Monitoring | Synapse