Efficient Methods for
Computational Imaging and Radiology
Sean I. Young, PhD
Instructor, Harvard Medical School
Research Affiliate, CSAIL, MIT
Apr 25, 2023
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 2
Outline
• Previous and Current Research
• Computational imaging and radiology
• Future research directions and conclusion
• Teaching and outreach
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 3
About me...
Stanford, CASydney, NSW Boston, MA
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 4
About me...
Stanford, CASydney, NSW
—PhD, EE
Boston, MA
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 5
Efficient Algorithms for Imaging and Vision Problems Sean I. Young Bernd Girod Slide 8 of 9
Select publications
Young et al. Graph Laplacian regularization for optical flow estimation IEEE Trans. Image Process., 2019
Input image pair
Pixel-graph construction
Baseline flow
Our flow field
Optical flow estimation as a graduated non-convex graph-regularized optimization problem. (Work done
during PhD at UNSW.)
Sydney, NSW
—PhD, EE
—Video Compression
—Inverse Problems
About me...
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 6
About me...
Efficient Algorithms for Imaging and Vision Problems Sean I. Young Bernd Girod Slide 4 of 9
Select publications
Young et al. Fast optical flow extraction from compressed video IEEE Trans. Image Process., 2020
Decoded picture
Decoded motion parameters
Recovered motion
Ground-truth motion
Extraction of sub-pixel-accurate motion fields from coded motion parameters using weighted filtering. We
achieve a 100× speed-up over estimating the motion from scratch.
Sydney, NSW
—PhD, EE
—Video Compression
—Inverse Problems
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 7
About me...
Efficient Algorithms for Imaging and Vision Problems Sean I. Young Bernd Girod Slide 6 of 9
Select publications
Young et al. Gaussian lifting for fast bilateral and nonlocal means filtering IEEE Trans. Image Process., 2020
Original picture
Filtered picture
Filter error (52.5dB)
Bilateral filtering posed as a pseudo-wavelet transform in the bilateral space. Up to a 5× speed-up over
other fast bilateral filter implementations.
Sydney, NSW
—PhD, EE
—Video Compression
—Inverse Problems
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 8
About me...
Stanford, CA
—Postdoc, EE
Sydney, NSW Boston, MA
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 9
Stanford, CA
—Postdoc, EE
—Compression
—Computational Imaging
About me...
Efficient Algorithms for Imaging and Vision Problems Sean I. Young Bernd Girod Slide 3 of 9
Select publications
Young et al. NLOS surface reconstruction using the Directional LCT CVPR (oral), 2020
Capture setup
Captured transients
Recovered surface
Ground-truth surface
Estimation of hidden surface normals posed as a vectorial deconvolution problem. We achieve a 1,000×
speed-up over previous methods.
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 10
About me...
Efficient Algorithms for Imaging and Vision Problems Sean I. Young Bernd Girod Slide 4 of 9
Select publications
Young et al. Fast optical flow extraction from compressed video IEEE Trans. Image Process., 2020
Decoded picture
Decoded motion parameters
Recovered motion
Ground-truth motion
Extraction of sub-pixel-accurate motion fields from coded motion parameters using weighted filtering. We
achieve a 100× speed-up over estimating the motion from scratch.
Stanford, CA
—Postdoc, EE
—Compression
—Computational Imaging
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 11
Stanford, CA
—Postdoc, EE
—Compression
—Computational Imaging
About me...
Efficient Algorithms for Imaging and Vision Problems Sean I. Young Bernd Girod Slide 5 of 9
Select publications
Young et al. Transform quantization for CNN compression IEEE Trans. Pattern Anal. Mach. Intell., 2020
Kernel quantization and pruning posed as a rate-distortion optimization problem. We achieve a 10×
inference speed-up over the default 8-bit quantization.
2021
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 12
About me...
Boston, MA
—Instructor, Radiology
Stanford, CASydney, NSW
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 13
Boston, MA
—Computational Radiology
—Computational Imaging
About me...
Young et al. Supervision by denoising. IEEE Tra n s. Pattern Anal Mach Intell, 2022
Semi-supervised medical image segmentation posed as a denoising problem. We achieve a 10x reduction in the
number of labeled examples required while significantly improving reconstruction accuracy.
FS reference nnU-Net Post-denoise Temporal ensembling SUD (ours)
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 14
About me...
Young et al. Fully convolutional SVR for single-stack MRI. CVPR, 2024
Slice-to-volume reconstruction as a self-supervised image registration task. We enable 3D reconstruction from a
single stack of MR slices, which has not previously been done.
Monocular depth estimation [4–8] Single-stack slice-to-volume reconstruction (Ours)
Boston, MA
—Computational Radiology
—Computational Imaging
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 15
About me...
Young et al. Fully convolutional SVR for single-stack MRI. CVPR, 2024
Slice-to-volume reconstruction as a self-supervised image registration task. We enable 3D reconstruction from a
single stack of MR slices, which has not previously been done.
Monocular depth estimation [4–8] Single-stack slice-to-volume reconstruction (Ours)
Boston, MA
—Computational Radiology
—Computational Imaging
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 16
Outline
• Introduction to myself and my research
• Computational imaging and radiology
• Future research directions and conclusion
• Teaching and outreach
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 17
Computational Radiology
Inverse problems
3D Object
Reflections
on a wall
Non-line-of-sight surface reconstruction
(Young, 2020)
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 18
Inverse problems
Computational Radiology
3D Volume
2D Acquisition
Slice-to-volume reconstruction (SVR)
(Young, 2024)
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 19
Efficient methods
Computational Radiology
Linear algebra
and signal processing
Signal processing
and machine learning
Machine learning
and signal compression
Transform quantization for CNN compression
(Young, 2022)
Efficient algorithms for computational radiology and imaging Sean I. Young Slide 20
Ongoing Work
3D brain imaging using conventional cameras and machine learning
Dissection photography is routine in nearly every brain bank. Dissected into coronal slices and photographed
before further blocking and histological slices. Leverage existing data for 3D reconstruction.
(a) Slab preparation (b) Slab photographs (c) 3D reconstruction