Efficient Neural Architecture Search on Low-Dimensional Data for OCT Image Segmentation
This addresses the computational bottleneck in NAS for medical imaging, offering a more efficient method for researchers and practitioners in this domain, though it is incremental as it builds on existing NAS techniques.
The paper tackled the problem of time-consuming neural architecture search (NAS) for medical imaging by proposing an efficient approach that searches on low-dimensional data and transfers to high-dimensional data, reducing search time by 87.5% for OCT image segmentation while maintaining similar performance.
Typically, deep learning architectures are handcrafted for their respective learning problem. As an alternative, neural architecture search (NAS) has been proposed where the architecture's structure is learned in an additional optimization step. For the medical imaging domain, this approach is very promising as there are diverse problems and imaging modalities that require architecture design. However, NAS is very time-consuming and medical learning problems often involve high-dimensional data with high computational requirements. We propose an efficient approach for NAS in the context of medical, image-based deep learning problems by searching for architectures on low-dimensional data which are subsequently transferred to high-dimensional data. For OCT-based layer segmentation, we demonstrate that a search on 1D data reduces search time by 87.5% compared to a search on 2D data while the final 2D models achieve similar performance.