Multi-view Analysis of Unregistered Medical Images Using Cross-View Transformers

Multi-view medical image analysis often depends on the combination of information from multiple views. However, differences in perspective or other forms misalignment can make it difficult to combine views effectively, as registration is not always possible. Without registration, only be combined at a global feature level, by joining vectors after pooling. We present novel cross-view transformer method transfer between unregistered level spatial maps. demonstrate this multi-view mammography and chest X-ray datasets. On both datasets, we find that links maps outperform baseline model joins