📝 AI4EO Map Your City Challenge 2024

---

Challenge Summary

Participants of this challenge are asked to develop a multi-modal model to estimate the construction year of any given building from 3 modality inputs: street-view imagery, VHR resolution top-view imagery, Medium Resolution S2 imagery. For half of the test set, street-view imagery will be missing, so the developed solution should address the issue of missing modality.

My solution includes modeling two classification models: type I where training and inference are conducted on all modalities, type II where training is done on all modalities, but inference is conducted only on top-view imageries. The second model is inspired by the Shared-Specific Feature Modelling approach in 3.

• Type I:

  • Feature Concatenation

    

  • Geometric Fusion

    

• Type II:

  

🎛 Development environment

---

mamba env create --file environment.yml
mamba activate ai4eo

💎 References

---

1. Model Fusion for Building Type Classification from Aerial and Street View Images
2. Multi-modal fusion of satellite and street-view images for urban village classification based on a dual-branch deep neural network
3. Multi-Modal Learning With Missing Modality via Shared-Specific Feature Modelling