Saved in:
Bibliographic Details
Main Authors: Ferede, Fisseha A., Khalighifar, Ali, John, Jaison, Venkataraman, Krishnan, Khairy, Khaled
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.07043
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917798788202496
author Ferede, Fisseha A.
Khalighifar, Ali
John, Jaison
Venkataraman, Krishnan
Khairy, Khaled
author_facet Ferede, Fisseha A.
Khalighifar, Ali
John, Jaison
Venkataraman, Krishnan
Khairy, Khaled
contents We propose a novel optical flow based approach to enhance the axial resolution of anisotropic 3D EM volumes to achieve isotropic 3D reconstruction. Assuming spatial continuity of 3D biological structures in well aligned EM volumes, we reasoned that optical flow estimation techniques, often applied for temporal resolution enhancement in videos, can be utilized. Pixel level motion is estimated between neighboring 2D slices along z, using spatial gradient flow estimates to interpolate and generate new 2D slices resulting in isotropic voxels. We leverage recent state-of-the-art learning methods for video frame interpolation and transfer learning techniques, and demonstrate the success of our approach on publicly available ultrastructure EM volumes.
format Preprint
id arxiv_https___arxiv_org_abs_2410_07043
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Z-upscaling: Optical Flow Guided Frame Interpolation for Isotropic Reconstruction of 3D EM Volumes
Ferede, Fisseha A.
Khalighifar, Ali
John, Jaison
Venkataraman, Krishnan
Khairy, Khaled
Image and Video Processing
Computer Vision and Pattern Recognition
We propose a novel optical flow based approach to enhance the axial resolution of anisotropic 3D EM volumes to achieve isotropic 3D reconstruction. Assuming spatial continuity of 3D biological structures in well aligned EM volumes, we reasoned that optical flow estimation techniques, often applied for temporal resolution enhancement in videos, can be utilized. Pixel level motion is estimated between neighboring 2D slices along z, using spatial gradient flow estimates to interpolate and generate new 2D slices resulting in isotropic voxels. We leverage recent state-of-the-art learning methods for video frame interpolation and transfer learning techniques, and demonstrate the success of our approach on publicly available ultrastructure EM volumes.
title Z-upscaling: Optical Flow Guided Frame Interpolation for Isotropic Reconstruction of 3D EM Volumes
topic Image and Video Processing
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2410.07043