IMAGE PROCESSING

PRESENTED BY DR. MABEL C. UGWUJA

HOD,DEPARTMENT OF RADIOGRAPY AND RADIATION SCIENCES
STATE UNIVERSITY OF MEDICAL AND APPLIED SCIENCES,IGBO-ENO,ENUGU STATE
IMAGE PROCESSING
OVERVIEW
 OUTLINES

• OBJECT
• IMAGE
• IMAGE QUALITY
• FACTORS THAT DEGRADE IMAGES
• ANALOGUE FILM PROCESSING STEPS
• DIGITAL FILM PROCESSING STEPS
 BASIC CONCEPTS IN IMAGING
OBJECT AND IMAGE
OBJECT
Physical or conceptual entity that existed in reality
• Property which is distributed in multidimension space.
• Can be measured .
• Denoted with symbol “f”

forms
• Physical =tangible ,exist in physical
• Mathematical= Abstract concept/entity==mathematical calculations eg nos ,vectors, or matrices.
• Programming= codes
• Artistic= creative art
• Philosophical =idea, concept, reality
• Cognitive=mental representation, thoughts, perceptions
WHAT IS AN IMAGE?

• Visual representation of anything, • Created through Various means such as :

such as object, person, scene, idea. 1. Photographic =cameras

2. Digital= using computer soft wares
3. Medical images=visualize body internal
• Visual depiction/representation of structures
an object. 4. Scientific= communicate scientific data
5. Graphic =visual communication
• Measurement that has been made 6. Artistic =by artists, painting,

that is the best demonstration of 7. Animated=animations, GIFs, etc.

8. Thermal=temperature differences(e.g.
object in distribution.
thermographic images, infrared images
9. Synthetic =Artificial intelligence
• Image has multi-dimension: 2D or 10. Satellite= from space
3D. Denoted with symbol “g” 11. 3D Images= Virtual reality
 RELATIONSHIP BETWEEN AN OBJECT AND
IMAGE

1. Representation: Image represents an object, conveying its visual properties

2. Abstraction : is an abstract version of the object , reducing its complexity to a visual
essence.
3. Perception: Image influences how an object is perceived and interpreted.
4. Communication: The Image can communicate information about the object to
others.
5. Symbolism: The image can carry symbolic meaning beyond the physical object itself
6. In a perfect world , the image looks exactly like an object, but in the real world, it is
impossible. It means that f ≠ g in real world.
IMAGE QUALITY
THE LEVEL OF CLARITY ,DETAILS AND OVER ALL
VISUAL APPEAL OF AN IMAGE
Measured by 2 Major indices; • FACTORS THAT AFFECT IMAGE QUALITY
• Subjective (can’t be measured or quantified, • EXPOSURE FACTORS
subject to observer’s experience, visual acuity
(KVP, MAs, exposure time , distance (SID,OID))
and viewing condition )
• Patient factors

• Physical indices(can be measured or (habitus , co-operation/motion)

quantified, for example; • Radiographic Equipment

Tube quality/age, generator, control unit,

• RESOLUTION collimator, image receptor quality(film or
• CONTRAST digital)

• NOISE
FACTORS THAT AFFECT IMAGE QUALITY
(CONT.)
• Image receptor factor(film type , • Quality control and assurance
Regular equipment tests and maintenance,
detector quality and film processing
quality controls, etc.
algorithms)
• Positioning and alignment In radiography the quality of an image is
described in terms of its;
• Artefact reduction • Resolution
• Contrast
Grid use and quality, Scatter radiation
• Noise level
reduction
• Artefact
• Image processing and • Sharpness
enhancement (processing algorithm , • Distortion
filtering and noise reduction) • Dynamic range
 RESOLUTION (ABILITY TO DISCRIMINATE B/W OR TO DETECT FINE
DETAILS AS SEPARATE ENTITIES)
ALSO (NO OF TINY DOTS (PIXELS) THAT MAKE UP AN IMAGE )
IT DETERMINES THE LEVEL OF CLARITY OF AN IMAGE

• It is the ability of an imaging system to differentiate 2 closely spaced

structures/entities.
• Temporal resolution

TYPES OF RESOLUTION the ability of an imaging system to

capture and display changes in an object
• Spatial resolution.
or scene over time. (measuring no of
This measures the no of pixels/length
frames per sec, mostly seen in video
(Ability of an imaging system to distinguish between 2 closely spaced images and animations) frame rate.
objects/points in space)It defines the level of detail that can be seen in an image.
Measured in units of distance. Can either be:

• Axial resolution or longitudinal resolution(Ability of an imaging system to • Colour resolution

distinguish 2 points along the longitudinal axis of an imaging beam).very Ability of am imaging system to
crucial in CT, MRI Imaging
accurately capture/display colours of the
• lateral resolution or transverse resolution, (Ability of an imaging system to
object. Measures the no of bits
distinguish 2 points along the transverse axis of the imaging beam
representing colour information.
 EFFECTS OF RESOLUTION ON AN IMAGE

• CLARITY(Clearer and more detailed) • CROPPING(high resolution can be cropped

aggressively without it loosing its clarity)
• SHARPNESS(sharper and better defined)
• PRINTING(can be printed without it loosing its
• DETAIL (captures more subtle details) quality)
• TEXTURE (high resolution images can • OVER ALL QUALITY(higher resolution images
capture more subtle textures ) have more professional look)

• NOISE(poor resolution=nosier/grainier A good imaging system should have a good

images) resolution system that handles both axial and lateral
resolution at the same time.
• ARTIFACTS (poor resolution =pixilation,
The higher the resolution, the more the image
blurriness)
produced by an imaging system is. The lower the
• ZOOMING (high resolution can be zoomed resolution, the less detailed and more pixelated the
without it loosing its clarity) images produced by the imaging system will be.
 CONTRAST
DIFFERENCE IN BRIGHTNESS, TONE OR COLOUR
B/W 2 OR MORE OBJECTS OR AREAS IN AN IMAGE

EFFECTS OF CONTRAST ON IMAGE • Detail and texture=can reveal or

• Visual hierarchy= Guides viewer’s attention hide details and texture,
• Mood and atmosphere=Can evoke the depending on the level of contrast.
viewer’s emotions and moods or
atmosphere.
• Noise and grain= Very high

• Depth and dimension=can create a sense

contrast make noise and grain
of depth or flatness very obvious
• Legibility= affects readability • Colour accuracy=very high
• Aesthetics =can enhance or distract from contrast can lead to colour clipping
over all visual appeal or loss of details.
TYPES OF CONTRAST

• High contrast (Large difference in

• INTRINSIC
brightness or density between
Natural differences in tissue density structures)
or composition of structures e.g. • Low contrast (small difference in
Air,Bone,fluid brightness or density between structures,
making difficult to differentiate them)
• EXTRINSIC
• Spatial contrast (difference in
Contrast agent introduced to enhance brightness or density between adjacent
the contrast, or an imaging technique structures
adopted to enhance contrast .e.g. fat • Temporal contrast (Difference in
suppression technique brightness or density between images
taken at different times)
FACTORS AFFECTING CONTRAST

• Imaging modality(Each imaging modality • Contrast agents affects contrast

have unique modality of operation,
by altering the tissue or proton
hence, unique in contrast of the images
produced).CT,MRI,U/S,X-RAY produce densities. E.g. barium ,iodine,
unique image gadolinium etc.
• Tissue density(differences in tissue
• Imaging processing techniques
densities affects the contrast on the
image produced) (image processing techniques
• Proton density(typically seen in MRI like windowing, levelling,
Images.it is the differences in proton filtering can enhance or reduce
densities of the closely related structures) contrast of an image)
CONTRAST RESOLUTION
-ABILITY OF AN IMAGING SYSTEM TO DISTINGUISH BETWEEN OBJECTS OR
FEATURES WITH DIFFERENT CONTRAST LEVELS

• IT’S A MEASURE OF HOW WELL TECHNIQUES TO IMPROVE

AN IMAGING SYSTEM CAN CONTRAST RESOLUTION
RESOLVE SUBTLE DIFFERENCES
IN CONTRAST
 NOISE
UNWANTED, CHAOTIC, RANDOM VARIATIONS OR FLUCTUATIONS IN THE
BACKGROUND OF AN IMAGE SIGNAL THAT CAN DEGRADE THE IMAGE
QUALITY

SOURCES OF NOISE IN IMAGING SYSTEM • Noise MANIFEST ON IMAGES AS ;

• Thermal noise= temperature fluctuations in the • GRAIN(Visible speckles or texture)
sensor
• SPECKLES (small white or black dots)
• Quantization noise= Caused by analogue to
digital conversion • SALT AND PEPPR NOISE(random black
or white pixels)
• Electronic noise=from either filters or electronic
components • COLOUR NOISE ( random variations in
• Readout noise=introduced during image colour)
readout. • DIGITAL ARTEFACTS (unwanted
• Sensor noise= from image sensor patterns or distortions
NOISE REDUCTION TECHNIQUES

• Image Filtering(mean , median, • Improved detector technology (quantum dot

Gaussian , wiener, wavelet Denoising, technology, multiple-pixel architectures ,
bilateral filter, machine learning-based advanced material to improve sensitivity
and reduce noise)
filter) are applied through various
domains to remove noise from the • Optimized imaging protocols(adjusting
various imaging system parameters can
image.
help much in noise result)
• Image processing algorithms(Gaussian
• Averaging multiple images(this is the
blur ,Median Filtering, bilateral simplest and an effective means of noise
filtering ) are some of the algorithm reduction).this apart from noise reduction ,
that can be applied to reduce noise it can improve signal to noise ratio and
enhance image details.
 IMAGE PROCESSING
NEED?

• To improve on the imperfections which are usually associated with the

images as a result of the instrument.
• To improved on the visual appearances of the image the human viewer.
• To prepare images for measurement of the features and the structures
present.
 LATENT IMAGE FORMATION
IMAGE PRODUCED ON X-RAY FILM AFTER EXPOSURE BUT PRIOR TO DEVELOPMENT

• FIRST stage in photochemical process

• Produced after exposure (produce changes within the silver halide
crystals, necessarily to render the crystals developable)
• Emulsion and silver bromide (suspension of AgBr)
• Radiation interacts with AgBr on exposure to form Ag + Br
• On development, Ag+ + e-(from developer) = Ag
• Gelatin on the emulsion of the film functions as a suspension and binder
agent for the silver bromide crystals.
ELECTRON TRAP

• Low energy/electron trap

• Shallow traps
• Deep traps
• Lattice defects and impurities
• Neutral colloids from impurities (2 EFFECTS)
1. Deepens electron trap, thereby increasing stability.
2. Defects in the crystal lattice
STEPS INVOLVED IN ANALOGUE FILM
PROCESSING

• EITHER MANUAL OR AUTOMATIC • FIXING

1. MANUAL METHOD Thiosulphate ions remove unexposed silver
• RECEIVING AND PREPARATION OF FILM halides, thus stabilizes the image.
FOR PROCESSING • WASHING
• DEVELOPMENT(exposed silver halide Film is washed to remove all the residual
crystals reacts with developer(reducing chemicals. No chemical reaction occurs
agent , hydroquinone , phenidone) to form here.
silver. Conversion of latent image to
visible image happens here) • DRYING

• RINSE(halts development and prevent • Air drying. Taken to a dryer cabinet with
over development) controlled temperature
AUTOMATIC FILM PROCESSING

Use of machine Disadvantages Advantages

High initial cost
• Film loader High cost of • Consistency
maintenance
• Developer tank and repairs • Increased efficiency
High cost of
• Fixer tank chemical • Reduced labour
consumption
• Wash tank and disposal • Improved image quality
Limited
• Dryer chamber • Faster processing time
flexibility for
• Controls special
processing
requirements
DIGITAL IMAGE PROCESSING: MANIPULATION AND ANALYSIS OF DIGITAL IMAGES TO ENHANCE,
RESTORE, OR EXTRACT INFORMATION.

It involves various techniques Techniques involves;

and algorithms to; • Filtering
Enhance • Transformation
Restore • Thresholding
Extract • Segmentation
Analyze • Feature extraction
• Image restoration
FILTERING TECHNIQUES IN DIGITAL IMAGE
PROCESSING
TYPES OF FILTERING • Sobel filtering(use of horizontal and vertical
gradient operators to detect edges and correct it)
• Low pass(removes high frequency noise)
• Laplacian filtering(use of second order derivatives
• High pass(enhances high frequency details and
to detect edge errors and correct it)
edges)
• Anisotropic filtering(smoothing of the image by
• Band- pass filtering(removes or enhances
adapting to local features)
specific range of frequencies)
• Bilateral filtering(smoothens the image while
• Median filtering(reduces salt and pepper noise
preserving the edges)
by pixel replacement)
The choice of filter depends on the application
• Gaussian filtering(using Gaussian distribution to
smoothen image) Each filter has its own strengths and
weaknesses.
• Wiener filtering(use of minimum mean square
error to smoothen and optimally reduce noise )
FILTERING IN DIGITAL IMAGE PROCESSING
(CONT.)
Why filtering? Image processing operations
• Noise reduction 1. Filtering
• Image smoothing 2. Transformation
• Edge detection 3. Morphological operations
• Image enhancement 4. Image reconstructions
5. Image display
6. Image storage
STEPS IN DIGITAL IMAGE PROCESSING

1. Image acquisition( Acquiring the • Image restoration(correcting

image from the source) degradation and all distortions )
2. Image import(transferring the • Image segmentation(dividing
image into a digital processor) the image into regions)
3. Image preprocessing(noise
• Feature extraction(identifying
reduction)
relevant features)
4. Image enhancement(adjusting
• Image analysis(interpreting and
brightness, contrast, colour
balance)
understanding image contents)
DICOM
(DIGITAL IMAGING AND COMMUNICATION IN MEDICINE)

DICOM
• Medical imaging
formats=supports various • Facilitates interoperability
imaging modalities • Ensures consistency
• Standard for handling ,storing, • Enable efficient storage and
printing and transmitting transmission of medical images.
information. • Widely in use in;
• File format definition and network • PACS
communications protocols
• RIS
• ELECTRONIC HEALTH RECORDS
THANK YOU FOR LISTENING