invariant feature transform (SIFT)33 and random sample consensus (RANSAC)34 may outperform the state-of-the-art DL methods for image matching35. We designed a general end-to-end DL framework for information extraction from CXR images (X-data)