Abstract:Objective. Auto-segmentation in mouse micro-CT enhances the efficiency and consistency of preclinical experiments but often struggles with low-native-contrast and morphologically complex organs, such as the spleen, resulting in poor segmentation performance. While CT contrast agents can improve organ conspicuity, their use complicates experimental protocols and reduces feasibility. We developed a 3D Cycle Generative Adversarial Network (CycleGAN) incorporating anatomy-constrained U-Net models to leverage contrast-enhanced CT (CECT) insights to improve unenhanced native CT (NACT) segmentation.Approach.We employed a standard CycleGAN with an anatomical loss function to synthesize virtual CECT images from unpaired NACT scans at two different resolutions. Prior to training, two U-Nets were trained to automatically segment six major organs in NACT and CECT datasets, respectively. These pretrained 3D U-Nets were integrated during the CycleGAN training, segmenting synthetic images, and comparing them against ground truth annotations. The compound loss within the CycleGAN maintained anatomical fidelity. Full image processing was achieved for low-resolution datasets, while high-resolution datasets employed a patch-based method due to GPU memory constraints. Automated segmentation was applied to original NACT and synthetic CECT scans to evaluate CycleGAN performance using the Dice Similarity Coefficient (DSC) and the 95th percentile Hausdorff Distance (HD95p).Main results.High-resolution scans showed improved auto-segmentation, with an average DSC increase from 0.728 to 0.773 and a reduced HD95p from 1.19 mm to 0.94 mm. Low-resolution scans benefited more from synthetic contrast, showing a DSC increase from 0.586 to 0.682 and an HD95preduction from 3.46 mm to 1.24 mm.Significance.Implementing CycleGAN to synthesize CECT scans substantially improved the visibility of the mouse spleen, leading to more precise auto-segmentation. This approach shows the potential in preclinical imaging studies where contrast agent use is impractical.

S-CycleGAN: A Novel Target Signature Segmentation Method for GPR Image Interpretation

Segment anything model for subgrade distress detection with GPR

A Gans-Based Deep Learning Framework for Automatic Subsurface Object Recognition From Ground Penetrating Radar Data

Dual-Task GPR Method: Improved Generative Adversarial Clutter Suppression Network and Adaptive Target Localization Algorithm in GPR Image

A GPR Resolution Enhancement Method Based on Weakly Supervised Learning

CycleGANs for Semi-Supervised Defects Segmentation

CycleGAN Based Clutter Suppression and Pipeline Positioning Method for GPR Image

CycleSeg: Simultaneous synthetic CT generation and unsupervised segmentation for MR‐only radiotherapy treatment planning of prostate cancer

A Novel Perceptual Constrained cycleGAN With Attention Mechanisms for Unsupervised MR‐to‐CT Synthesis

Structure Preserving Cycle-GAN for Unsupervised Medical Image Domain Adaptation

S-CycleGAN: Semantic Segmentation Enhanced CT-Ultrasound Image-to-Image Translation for Robotic Ultrasonography

Generation of Contrast-Enhanced CT with Residual Cycle-Consistent Generative Adversarial Network (Res-Cyclegan)

Integrated GANs: Semi-Supervised SAR Target Recognition

CycleSGAN: A cycle-consistent and semantics-preserving generative adversarial network for unpaired MR-to-CT image synthesis

CAM-CycleGAN: A Weakly Supervised Segmentation Method for Medical Images Based on Cycle Consistensy Generative Adversarial Networks

Seg-CycleGAN : SAR-to-optical image translation guided by a downstream task

Simulation of GPR B-Scan Data Based on Dense Generative Adversarial Network.

Anatomy-constrained synthesis for spleen segmentation improvement in unpaired mouse micro-CT scans with 3D CycleGAN

SUSAN: Segment Unannotated image Structure using Adversarial Network

CycleSegNet: Object Co-segmentation with Cycle Refinement and Region Correspondence

GPR B-Scan Image Augmentation Via GAN with Multiscale Discrimination Strategy