Search papers, labs, and topics across Lattice.
4
0
4
0
OmniRSCLIP achieves strong performance across heterogeneous remote sensing modalities while preserving the advantages of RGB-based models.
VLMs can be repurposed for multispectral and SAR tasks with minimal adaptation, achieving impressive accuracy without retraining from scratch.
Scene graphs are all you need: AeroRAG shows that structured knowledge retrieval from visual data significantly boosts LLM performance on fine-grained visual reasoning tasks.
Current facial expression editing models can't simultaneously preserve identity and accurately manipulate expressions, revealing a critical need for better fine-grained instruction following.
