incremental learning

This work visually analyzes anatomical variants of liver vessels anatomy to maximize similarity for finding suitable living donor-recipient pairs. We leverage incremental learning in a cascade feature mapping way via updating input CTA training model to optimize segmentation capability. A ternary-tree-based approach is proposed to map all possible liver vessel variants into their respective tree topologies. The ternary tree in-order traversing is designed to efficiently compare the digital strings of two anatomically varied vessel structures to find a suitable match. Experiments through visual illustrations and quantitative analysis demonstrated our method computed very efficiently for finer visualization of liver tree structures. read more