Comparing feature fusion strategies for Deep Learning-based kidney stone identification
This work addresses kidney stone identification for urologists, offering an incremental improvement in diagnostic accuracy.
The paper tackled the problem of kidney stone classification by developing a deep-learning method that fuses image features from multiple viewpoints, improving classification precision by over 10% compared to single-view models.
This contribution presents a deep-learning method for extracting and fusing image information acquired from different viewpoints with the aim to produce more discriminant object features. Our approach was specifically designed to mimic the morpho-constitutional analysis used by urologists to visually classify kidney stones by inspecting the sections and surfaces of their fragments. Deep feature fusion strategies improved the results of single view extraction backbone models by more than 10\% in terms of precision of the kidney stones classification.