Cross-Dataset Multiple Organ Segmentation From CT Imagery Using FBP-Derived Domain Adaptation

Abstract

Multi-organ segmentation from whole-body computed tomography (CT) scans has gained increasing research interest over recent years. While the learning-based segmentation algorithm has lately achieved tremendous success, the need for detailed annotation of multiple organs further increases the manual burden. With a limited number of annotated volumetric datasets, it would be beneficial to apply the trained model from such a set to CT images acquired from other sites with different scanners. Nevertheless, the discrepancy among training and testing images significantly deteriorates segmentation performance. While there are many domain adaptation efforts, in this work we proposed a filtered back-projection based algorithm for performing domain adaptation for CT imagery. An optimal CT reconstruction kernel was obtained by minimizing the disparity between two images. Furthermore, since the Gaussian kernel is an eigen-function of the Fourier transformation, the adaptation computation was proven to be simple linear filtering. The proposed method was tested and compared with multiple methods to demonstrate improvement by employing such a model/theory-based adaptation approach. The proposed method, used in conjunction with a common convolutional neural network, such as the U-Net or V-Net, with or without the domain adaptation, achieves high accuracy in a multiple-organ segmentation task. Approximately 30% of data was used for training, 70% was used for testing, and an average dice of 0.88 was achieved in 8 organs.

Three-Dimensional Digital Reconstruction of the Cerebellar Cortex: Lobule Thickness, Surface Area Measurements, and Layer Architecture

“…The laminar structure of the cerebellum, namely, the granular cell layer (including the single cell Purkinje cell layer) and the molecular layer, was segmented using the semi-automatic software 3D Slicer [ 42 ] and a 2D convolutional neural network (CNN) based approach [ 43 ].…”

“…3 Segmentation of cerebellar layers. The white matter (WM) and granular layer were segmented by an interactive method [42] and the pial surface was segmented by automatic deep learningbased method [43], see "Materials and Methods" section for detail. It should be noted that while this finding is consistent with data derived from local histological analyses [21], histology-based approaches have typically been unable to map the laminar thickness across the entire cerebellum.…”

“…Fig.3Segmentation of cerebellar layers. The white matter (WM) and granular layer were segmented by an interactive method[42] and the pial surface was segmented by automatic deep learningbased method[43], see "Materials and Methods" section for detail.A. Coronal, high-resolution histological sections of the BigBrain cerebellum.…”

Multi-organ segmentation: a progressive exploration of learning paradigms under scarce annotation

“…Apart from the adversarial-based approach, Ma et al proposed a neural style transfer approach that reduces the data inconsistency between sources to minimize the variation in image attributes (including brightness, contrast, texture, etc) and integrates global information using atrous convolution and pyramidal pooling (Ma et al 2019). Huang et al enforced filtered back projection (FBP)-driven domain adaptation techniques to minimize the gap across different CT image datasets (Huang et al 2021), and fused the contextual information through global organs' registration. Fu et al employed the relatively fixed spatial location between organs as potentially transferable shared knowledge for domain adaptation of multi-center data (Fu et al 2020).…”