Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

Biomedical Imaging and Instrumentation

Biomedical Imaging and Instrumentation - Poster Session A

Poster F3 - 3-D Surface Reconstruction of Quasi-Cylindrical Structures Using Panoramic Digital Image Correlation

Thursday, October 24, 2024

10:00 AM – 11:00 AM EST

Location: Exhibit Hall E, F & G

Presenting Author(s)

Alec Mercer

Student - University of New Hampshire
University of New Hampshire
Sanford, Maine, United States

Co-Author(s)

Brendan K. Otani (he/him/his)

PhD Student
Washington University in St. Louis
St. Louis, Missouri, United States

Primary Investigator(s)

Matthew R. Bersi, PhD (he/him/his)

Assistant Professor in Mechanical Engineering and Materials Science
Washington University in St. Louis, United States

Introduction: Vascular disorders, such as aortic aneurysm and rupture, are dependent on the biomechanical properties of vasculature and supporting structures. Localized defects in the aortic wall have the potential to be catastrophic due to local changes in wall composition and structural integrity. Understanding the localized impact of mechanical defects on vascular microstructures can promote improved diagnosis and prognosis of vascular disorders.

To investigate the effects of localized mechanical defects, novel systems capable of capturing local deformations in real time on ex-vivo tissue structures are required. To fulfill this need, a panoramic digital image correlation system (pDIC) was developed to capture the full-field (entire surface) geometry of quasi-cylindrical structures. The pDIC system has the advantage of imaging the entirety of a structure without the need for recalibration after each capture, making dynamic captures of vessel pressurization over time possible.

Advancements in pDIC hardware and software workflows have enabled full-field, panoramic capture of quasi-cylindrical structures and reconstruction of surface geometry via image correlations using a static four-camera view. By using a tubular structure with known size and shape, such as a blunt hypodermic needle, this local biomechanical testing system and technique can be validated by first reconstructing the needle shape. Prior success of this reconstruction has established this test case as a reliable ground truth measurement. Future use and improvements on this novel mechanical testing system can improve our current understanding of vascular biomechanics and advance diagnosis of vascular conditions to reduce the impact of localized defects on catastrophic vascular failure.

Materials and

Methods: The pDIC system consists of four cameras organized such that two stereo views are achieved when a sample is placed vertically within a concave conical mirror (along its axis) (Fig.1). To successfully reconstruct sample geometries by pDIC, a speckle pattern must first be applied to the sample’s surface. Speckling is applied via airbrush wherein black ink is used to coat the sample prior to application of quick bursts of white ink to create small white dots on the surface of the sample. These small white dots are recognizable features for the image analysis software which are crucial to successful feature tracking and shape reconstruction.

The precision conical mirror reflects the surface speckle pattern, thus allowing the cameras to capture its unique features. Each camera captures an image of the sample that is unwrapped (converted from polar to cartesian coordinates) to be processed. Features conserved across images are first recognized using the Scale-Invariant Feature Transform (SIFT) method. Upon feature recognition and matching, a map of disparities between the reference and target image is generated. This map dictates how features in the target image must be morphed to become increasingly similar to the reference image. Next, the morphed images are correlated to the reference image and custom software (Matlab) uses the correlation results to map points back to the original image using the reference frame of the conical mirror and ultimately projects them into a coherent structure based on point-mapping parameters which replicate the surface geometry of the imaged structure.

Results, Conclusions, and Discussions: Experimental reconstructions of hypodermic needles are a useful approach to validate surface reconstructions and establish pDIC as an approach to measure local properties necessary to expand our understanding of quasi-cylindrical biological tissues. Successful needle reconstructions are a ground truth for the accuracy of the system and demonstrate feasibility of applying the method to similarly-sized biological samples (Fig.2). Implementation of robust feature-recognition algorithms have increased 3D reconstruction accuracy from the pDIC system while maintaining computational efficiency. Prior iterations have suffered from inaccurate surface geometry reconstructions due to differences in image perspective, while the current software workflow accommodates such disparities through pointwise correlation and a robust image morphing scheme. Additionally, current advancements in the software and image analysis procedure remove excess user intervention through automatic feature detection, disparity mapping, and surface generation. These advancements, combined with the current ground truth needle experiment, have contributed to increasing the reliability of the pDIC system for accurate 3D reconstructions of deformation in quasi-cylindrical biological tissues.

pDIC has the unique advantage of being a low-cost imaging system that can be deployed with relative ease. Additionally, pDIC is a method of non-contact optical assessment that can track local surface deformation, thus providing potential for ex-vivo analysis of complex biological tissues. The system can be easily scaled to accommodate samples of varying sizes to promote the study of murine ( < 2.5mm diameter) to human ( < 3.0 cm) vasculature.

Following the successful application of this updated workflow to 3D reconstruction of biological tissue (Fig.3), pDIC can be utilized in tandem with other imaging and histological methods to characterize localized defects in murine vascular structures. Given the near real-time acquisition, damaged aortae can be pressurized in the pDIC system until rupture. The pDIC can dynamically capture surface deformations to characterize the biomechanical properties around vascular defects to advance the current knowledge of rupture mechanics and the localized nature of aortic rupture. Building upon the updated analysis framework and validation presented in the current study, further advancements in this space will lead step toward spatiotemporal rupture prediction, understanding drivers of rupture, and improving diagnosis and prognosis of vascular conditions associated with rupture.

Acknowledgements (Optional): This work was made possible by support from the WashU Summer Engineering Fellowship (WUSEF) program as well as funding from the National Science Foundation (CAREER BMMB-2430666 to MRB)