3D IMAGE SEGMENTATION WITH FOURIER NEURAL OPERATOR
The computational complexity of 3D biomedical image segmentation often leads to out-of-memory errors in deep learning. To mitigate this, training with downsampled images is a common practice. However, this can lead to reduced accuracy when applying the model on original resolution images due to sensitivity to variations in image resolution. FNOSeg3D is introduced as a 3D segmentation model designed to be robust to variations in training image resolution. It leverages the Fourier neural operator (FNO) [1], a deep learning framework for learning mappings between functions in partial differential equations. FNO has the ability to enhance image resolution without explicit training on high-resolution data. This approach appears to be a promising advancement in the field of 3D biomedical image segmentation, particularly in terms of addressing the challenges related to image resolution variations during training.
The public datasets are available: https://cremi.org/data/, https://openorganelle.janelia.org/