Fig. 1: Overview of ELD framework.
From: Spatial landmark detection and tissue registration with deep learning
Proposed workflow of acquiring spatial landmarks and aligning sections within a CCF. The process can be divided into steps as follows. a, Obtain two different sections, A and B, for which we want to identify landmarks. b,c, Use a trained unsupervised spatial landmark detection network (b) or manually annotate the sections to obtain landmarks for both sections (c). d, After acquiring landmarks for both samples, register Section B to Section A using various landmark-based alignment methods, such as TPS or homography. e, Obtain the registered sample, which aligns Section B with Section A. f, Map the samples to a CCF, allowing cross-section comparisons and analysis.
