Comparison of 2D vs. 3D Unet Organ Segmentation in abdominal 3D CT images

Abstract

A two-step concept for 3D segmentation on 5 abdominal organs inside volumetric CT images is presented. Firsteach relevant organ’s volume of interest is extracted as bounding box. The extracted volume acts as input for asecond stage, wherein two compared U-Nets with different architectural dimensions re-construct an organ segmen-tation as label mask. In this work, we focus on comparing 2D U-Nets vs. 3D U-Net counterparts. Our initial resultsindicate Dice improvements of about 6% at maximum. In this study to our surprise, liver and kidneys for instancewere tackled significantly better using the faster and GPU-memory saving 2D U-Nets. For other abdominal keyorgans, there were no significant differences, but we observe highly significant advantages for the 2D U-Net interms of GPU computational efforts for all organs under study.

Comparing 3D, 2.5D, and 2D Approaches to Brain Image Segmentation

“…Our study extends the prior literature 10,12,13,32,33 in key ways. We provide the first comprehensive benchmarking of 3D, 2.5D, and 2D auto-segmentation approaches on brain MRIs measuring both accuracy and computational efficiency.…”

“…Compared to two-dimensional auto-segmentation tasks, the three-dimensional (3D) nature of brain MRIs makes auto-segmentation considerably more challenging (Figure 1). There have been three proposed approaches to handling auto-segmentation of 3D images: 1) analyze and segment one two-dimensional slice of the volume individually one at a time (2D), 10 2) analyze five consecutive two-dimensional slices at a time generate a segmentation the middle slice (2.5D), 11 and 3) analyze and segment the image volume in three-dimensional space (3D). 10 Although each method has shown some promise in medical image segmentation, a comprehensive comparison and benchmarking of these approaches for auto-segmentation of brain MRIs is lacking.…”

“…There have been three proposed approaches to handling auto-segmentation of 3D images: 1) analyze and segment one two-dimensional slice of the volume individually one at a time (2D), 10 2) analyze five consecutive two-dimensional slices at a time generate a segmentation the middle slice (2.5D), 11 and 3) analyze and segment the image volume in three-dimensional space (3D). 10 Although each method has shown some promise in medical image segmentation, a comprehensive comparison and benchmarking of these approaches for auto-segmentation of brain MRIs is lacking. Prior studies on comparing these auto-segmentation approaches have often not evaluated efficacy brain MRIs or limited their comparison narrowly to one deep learning architecture.…”

“…Prior studies on comparing these auto-segmentation approaches have often not evaluated efficacy brain MRIs or limited their comparison narrowly to one deep learning architecture. 10,12–14 Additionally, previous studies have focused primarily on segmentation accuracy and failed to evaluate more practical metrics such as computational performance and efficacy in data limited settings. As a result, it is difficult for clinicians and researchers to easily choose the appropriate auto-segmentation method for a desired clinical task.…”

Accelerating a Morphology-Preserving Adsorption Model by Deep Learning

“…adjacent slices). However, in the case of the U-Net-like architecture, many works [13,14] have reported that both 2D and 3D approaches lead to close results in the context of semantic segmentation for the used medical datasets. To address the issue of missing inter-slice information and bridge the gap further between 2D and 3D strategies, a distance transform was applied to the initial volume before slicing.…”