An Educated Warm Start For Deep Image Prior-Based Micro CT Reconstruction
This work addresses computational inefficiency in DIP-based reconstruction for micro CT imaging, offering a practical improvement for researchers in medical or biological imaging, though it is incremental as it builds on existing DIP methods.
The paper tackles the slow convergence of Deep Image Prior (DIP) for unsupervised image reconstruction by introducing a two-stage learning paradigm with supervised pretraining on simulated data, which speeds up and stabilizes reconstruction on real micro CT data of biological specimens, achieving faster convergence without specifying exact numbers.
Deep image prior (DIP) was recently introduced as an effective unsupervised approach for image restoration tasks. DIP represents the image to be recovered as the output of a deep convolutional neural network, and learns the network's parameters such that the output matches the corrupted observation. Despite its impressive reconstructive properties, the approach is slow when compared to supervisedly learned, or traditional reconstruction techniques. To address the computational challenge, we bestow DIP with a two-stage learning paradigm: (i) perform a supervised pretraining of the network on a simulated dataset; (ii) fine-tune the network's parameters to adapt to the target reconstruction task. We provide a thorough empirical analysis to shed insights into the impacts of pretraining in the context of image reconstruction. We showcase that pretraining considerably speeds up and stabilizes the subsequent reconstruction task from real-measured 2D and 3D micro computed tomography data of biological specimens. The code and additional experimental materials are available at https://educateddip.github.io/docs.educated_deep_image_prior/.