AI for Medical Imaging - Slides Summary
Lecture 1: Introduction
Medical Imaging
● Some info
○ Number of patents: ~ 3 million (Dec 2016), still growing
○ Bar is too high to go from Bench to Bedside
○ Major emphasis is on non-invasive procedures
○ Indispensable in the clinic
● Goal
○ Visualization of lesions and/or disease
○ Radiological interpretation
○ Rule-out or rule-in a diagnosis
Medical imaging evolution
● Old” = 2D; X-ray Radiography, nuclear
● “Recent” = 3D; CT, MRI, PET, etc.
● “New” = 4D; dynamic CT, MRI, US
● “Horizon” = 5D; 4D + function
Major advances in medical imaging
● Faster (dynamic)
● 4D protocols
● Farther (3D anatomy)
● Whole body imaging in less than 20 sec
● Functional (physiology)
● PET, SPECT, Optical, and MRS
● Fusion (anatomy & function)
● PET/CT, PET/MR, MR/CT, Optical/MR
Medical Imaging Modalities: Comparison
1
,AI for Medical Imaging - Slides Summary
Medical Images for the Physicians view
● Medical images are obtained using 3D/4D Volume Image Acquisition Systems
Advanced Image Processing is used in Medical Image Applications
● Enhancement
○ improves the visual quality of image, helps the clinician in his decision and thus
saves the patients' lives
○ More detailed image than before, e.g.,
2
,AI for Medical Imaging - Slides Summary
■ Enhancement of pulmonary structures
■ Correction of MR field bias
■ Restoration of lung epithelial cells
● (Volume) Rendering
○ Volume rendering represents a collection of methods used in computer graphics
and scientific visualization to create a 2D projection from a discretely sampled 3D
data set
○ An example of a 3D data set is a collection of MRI, CT, or MicroCT scanner 2D
slice images
○ For instance, a series of 2D slice images of a human brain can be assembled to
render 3D volume rendered images using a volume rendering algorithm
○ We stack 2D images to obtain a 3D image
○ We also have realistic tissue rendering
● Segmentation
○ Partitioning a digital image into multiple image segments/regions or image
objects (sets of pixels)
○ Goal is to simplify and/or change the representation of an image into something
that is more meaningful and easier to analyze
○ Typically used to locate objects and boundaries (lines, curves, etc.) in images
○ The result of image segmentation is a set of segments that collectively cover the
entire image, or a set of contours extracted from the image
● Modeling
● Fusion/registration
○ Process of spatially aligning two or more image datasets of the same scene
taken at different times, from different viewpoints, and/or by different sensors
○ Aligning two images of different sorts of the same patient (data fusion)
Summary
● Impact of recent advances in Medical Imaging
● Improved outcomes and yield
3
, AI for Medical Imaging - Slides Summary
● Reduced risk and morbidity
● Less procedure time, faster recovery
● Higher throughput and less cost
Medical Image Components
● Pixel depth
○ Number of bits used to encode the information of each pixel
● Photometric interpretation
○ Specifies how the pixel data should be interpreted for the correct image display
as a monochrome or color image
○ Clinical radiological images, like x-ray computed tomography (CT) and magnetic
resonance (MR) images have a gray scale photometric interpretation
○ Nuclear medicine images, like positron emission tomography (PET) and single
photon emission tomography (SPECT), are typically displayed with a color map
or color palette
● Metadata
○ Information that describes the image. It can seem strange, but in any file format,
there is always information associated with the image beyond the pixel data
○ Typically stored at the beginning of the file as a header and contains at least the
image matrix dimensions, the spatial resolution, the pixel depth, and the
photometric interpretation
● Pixel data
○ Section where the numerical values of the pixels are stored
○ According to the data type, pixel data are stored as integers or floating-point
numbers using the minimum number of bytes required to represent the values
Radiology Workflow
1. New patient arrives at hospital for radiological examination (exam).
2. Patient registers in Radiology area. If the patient is new, patient is registered in the
Hospital Information System (HIS).
3. Exam ordered at Radiology Information System (RIS) upon arrival at Radiology
registration desk. Exam accession number is automatically assigned, and a requisition is
printed
4
Lecture 1: Introduction
Medical Imaging
● Some info
○ Number of patents: ~ 3 million (Dec 2016), still growing
○ Bar is too high to go from Bench to Bedside
○ Major emphasis is on non-invasive procedures
○ Indispensable in the clinic
● Goal
○ Visualization of lesions and/or disease
○ Radiological interpretation
○ Rule-out or rule-in a diagnosis
Medical imaging evolution
● Old” = 2D; X-ray Radiography, nuclear
● “Recent” = 3D; CT, MRI, PET, etc.
● “New” = 4D; dynamic CT, MRI, US
● “Horizon” = 5D; 4D + function
Major advances in medical imaging
● Faster (dynamic)
● 4D protocols
● Farther (3D anatomy)
● Whole body imaging in less than 20 sec
● Functional (physiology)
● PET, SPECT, Optical, and MRS
● Fusion (anatomy & function)
● PET/CT, PET/MR, MR/CT, Optical/MR
Medical Imaging Modalities: Comparison
1
,AI for Medical Imaging - Slides Summary
Medical Images for the Physicians view
● Medical images are obtained using 3D/4D Volume Image Acquisition Systems
Advanced Image Processing is used in Medical Image Applications
● Enhancement
○ improves the visual quality of image, helps the clinician in his decision and thus
saves the patients' lives
○ More detailed image than before, e.g.,
2
,AI for Medical Imaging - Slides Summary
■ Enhancement of pulmonary structures
■ Correction of MR field bias
■ Restoration of lung epithelial cells
● (Volume) Rendering
○ Volume rendering represents a collection of methods used in computer graphics
and scientific visualization to create a 2D projection from a discretely sampled 3D
data set
○ An example of a 3D data set is a collection of MRI, CT, or MicroCT scanner 2D
slice images
○ For instance, a series of 2D slice images of a human brain can be assembled to
render 3D volume rendered images using a volume rendering algorithm
○ We stack 2D images to obtain a 3D image
○ We also have realistic tissue rendering
● Segmentation
○ Partitioning a digital image into multiple image segments/regions or image
objects (sets of pixels)
○ Goal is to simplify and/or change the representation of an image into something
that is more meaningful and easier to analyze
○ Typically used to locate objects and boundaries (lines, curves, etc.) in images
○ The result of image segmentation is a set of segments that collectively cover the
entire image, or a set of contours extracted from the image
● Modeling
● Fusion/registration
○ Process of spatially aligning two or more image datasets of the same scene
taken at different times, from different viewpoints, and/or by different sensors
○ Aligning two images of different sorts of the same patient (data fusion)
Summary
● Impact of recent advances in Medical Imaging
● Improved outcomes and yield
3
, AI for Medical Imaging - Slides Summary
● Reduced risk and morbidity
● Less procedure time, faster recovery
● Higher throughput and less cost
Medical Image Components
● Pixel depth
○ Number of bits used to encode the information of each pixel
● Photometric interpretation
○ Specifies how the pixel data should be interpreted for the correct image display
as a monochrome or color image
○ Clinical radiological images, like x-ray computed tomography (CT) and magnetic
resonance (MR) images have a gray scale photometric interpretation
○ Nuclear medicine images, like positron emission tomography (PET) and single
photon emission tomography (SPECT), are typically displayed with a color map
or color palette
● Metadata
○ Information that describes the image. It can seem strange, but in any file format,
there is always information associated with the image beyond the pixel data
○ Typically stored at the beginning of the file as a header and contains at least the
image matrix dimensions, the spatial resolution, the pixel depth, and the
photometric interpretation
● Pixel data
○ Section where the numerical values of the pixels are stored
○ According to the data type, pixel data are stored as integers or floating-point
numbers using the minimum number of bytes required to represent the values
Radiology Workflow
1. New patient arrives at hospital for radiological examination (exam).
2. Patient registers in Radiology area. If the patient is new, patient is registered in the
Hospital Information System (HIS).
3. Exam ordered at Radiology Information System (RIS) upon arrival at Radiology
registration desk. Exam accession number is automatically assigned, and a requisition is
printed
4
