A rigorous and efficient GPU implementation of level-set sparse field algorithm

Abstract

International audienceLevel-set methods have proven to be powerful and flexible tools in computer vision and medical imaging. Unfortunately, the flexibility of such models has historically resulted in long computational times and therefore limited clinical utility. In this context, we propose the first rigorous GPU implementation of the sparse field algorithm. We show that this model is able to reach high computational efficiency with no reduction in segmentation accuracy compared to its sequential counter-part

Fast level-set based image segmentation using coherent propagation

“…Significant speed-up was reported in a number of studies on implementing level-set on GPU. 9,19,20 However, in our experience, the performance of most optimized algorithms on 3D medical data is still not adequate for clinical practice. For example, typical 3D brain segmentation (256 3 ) takes more than 5 min on an ordinary PC and about 2 min on a GPU, when using the sparse field approach.…”

Study, Modelling and Implementation of the Level Set Method Used in Micromachining Processes

“…These data structures are not easily adapted to run on a parallel environment, thus, only a few parallel GPU-based implementations have been recently published and all of them in the field of medical image segmentation [12,[285][286][287]. Galluzzo et al proposed an implementation which stores the whole grid and emulates the SFM lists with one-dimensional arrays [287]. On the other hand, approaches that reduce memory storage by splitting the grid into blocks such that, only those blocks close to the front are stored in GPU memory have been presented [285,286].…”