📢 New Publication from GEAR Lab in Computers, Environment and Urban Systems!

We are thrilled to share that Yifan Yang (lead author) and Dr. Lei Zou, along with Bing Zhou, Daoyang Li, Binbin Lin, Joynal Abedin, and Mingzheng Yang, have published a new paper:

🎯 "Hyperlocal disaster damage assessment using bi-temporal street-view imagery and pre-trained vision models"

Fig. Pre- and post-hurricane street view comparison for disaster impact assessment

🔍 Study Highlights:

✅ Novel dual-channel approach combining pre- and post-disaster street-view images for hyperlocal damage assessment.

✅ Dataset: 2,249 pre/post-disaster image pairs, labeled into three damage categories.

✅ Accuracy boost: From 66.14% (Swin Transformer) to 77.11% with a dual-channel Feature-Fusion ConvNeXt model.

✅ Better change detection: Incorporating pre-disaster imagery helps detect damage changes more effectively.

✅ Practical impact: Supports decision-making in disaster response and resilience planning.

📄 Read the Full Paper

🎉 Congratulations to Yifan and the team for advancing GeoAI-driven disaster resilience research! 🚀