MedIX Summer 07 Lucia Dettori (room 745) [email protected]

MedIX  Summer 07 Lucia Dettori (room 745) ldettori@cti.depaul.edu

MedIX Summer 07 Lucia Dettori (room 745) [email protected] Projects Contrast Enhancement A brief summary of what has been done Things I would like to explore next Texture classification Evaluations of segmentation algorithms

Broad goal The big picture Manipulate (medical) images to facilitate the radiologists job of recognizing features and pathologies in radiological images Improve the visual quality of an image and automatically highlight certain features Give them a way to focus on subsets of the image that are of interest to them Contrast is all we see Human eye identifies details by contrasting an object (foreground) and its

background Improve the quality of the image by creating (color) contrast In our case we are talking about CT scan images with different levels of grey Contrast enhancement Take the gray level intensities of an image and

proportionally redistribute them Some mapping is necessary anyway since the images are based on 12 bits of information (gray levels ranges from 0 to 4095) and on these monitors we can only display 8 bits (gray levels from 0 to 255) Can we do a better job? Example Techniques implemented Linear binning Linearly

redistribute the intensities from a range of 0 - 4095 to a range of 0-255 over a chosen number of bins Nonlinear binning First identify clusters of intensities then use those to guide the redistribution of gray levels

Histogram equalization Soft tissues or bones? Radiologist might be interested in only some parts of the image: Soft tissues Lungs Bones Each of these correspond to a different

range of grey levels Local contrast enhancement Instead of trying to enhance the entire picture, concentrate the enhancing power in the range of intensities you are interested in Window enhancement Multiple windows enhancement Where are we now? Previous students have created an

application (C#) that allows the user to select windows and technique and display the enhanced image Things Id like to explore Find an objective way to measure the improvement resulting from the contrast enhancement See how these techniques perform when using different image modalities beyond CT-scans Explore additional contrast enhancement techniques

Projects Contrast Enhancement Texture classification A brief summary of what has been done Things I would like to explore next Evaluations of segmentation algorithms The big pictures Given a pre-segmented organ region, can

you tell me what it is: kidney, heart etc? It depends on its texture Identify image features that give texture information Find rules that distinguish the texture features of one organ from another Texture Classification Process at a glance Physician annotated Organ/Tissue Liver Kidney Bone

Spleen Apply filter To the image Texture Descriptors Heart Classification rules for tissue/organs in CT images

Classifier (Decision Tree) Image Data Set Organ Qty Image Organ Qty Kidney

223 Aorta 66 Liver 260 IP Fat 59

Spleen 95 Muscle 198 Trabecular Bone 39 SQ Fat 157

15 Total Images Lung 111 2 Image Step3 Texture features extraction Physician annotated

Organ/Tissue Liver Apply Gabor filters to the image Texture Descriptors Array of texture descriptors [T1, T2, T3, ,

Tn] For example: Mean, standard deviation, energy, entropy etc.. Step4 - Classification Physician annotated Organ/Tissue Classification performance measures

Apply Gabor filters to the image Classification rules for tissue/organs in CT images The process of identifying a region as part of a class (organ) based on its texture properties. Texture Descriptors

Decision tree Predicts the organ from the values of the texture descriptors Training / Testing Step5 Evaluating the classifier Things I would like to explore Physician annotated Organ/Tissue

Test Gabor texture descriptors on additional images and natural images Performance measures Apply Gabor filters to the image Texture Descriptors Wedgelet transfors

Fractal Dimensions Classification rules for tissue/organs in CT images Decision tree Projects Contrast enhancement Texture classification Evaluations of segmentation algorithms

Brief summary of what has been done Things I would like to explore next Texture segmentation Given an image, can you tell me how many organs you have? That was easy enough. Can you tell which organs they are? Identifying regions with similar texture

Identifying which texture it is to label the organ A couple of key questions Can you do it better by varying a parameter? How do you choose the values of your segmentation parameters? If it looks better is it really better? A couple of key questions

Parameter optimization Performance evaluation 1 0.87 3 0.56 4

2 0.50 0.75 Ground Truth Regions key Machine Segmentations

Increasing value of a segmentation parameter How do I decide what the optimal value of the parameter is? How good a segmentation is it? The goodness metric A single value that assigns a rating to a particular segmentation based on how well the machine segmented regions match

the regions in the ground truth images Region Categories Ground Truth vs. Machine Segmented Correctly Detected Over Segmented GT Under Segmented Missed MS Noise OVER SEGMENTED CORRECTLY

DETECTED Index for each region UNDER SEGMENTED A Missed region is a GT region that does not participate in any instance of CD, OS, or US A Noise region is an MS region that does not participate in any instance of CD, OS, or US

The Goodness Metric good = Correct Detection Index bad = 1-Correct Detection Index goodness = good-bad*weight 1.0 Ceiling = CDind Weight Range = CDind-1 Floor = 2*CDind-1 -1.0 How can we use the metric?

Create a set of ground truth mosaic using radiologist-labels images of pure patches of organ tissues Apply segmentation algorithm Optimize the segmentation parameters using the metric Apply optimized algorithm to the real image Ground Truth

T=1000; GM= - .94 T=2000; GM= - .02 T=4000; GM= .74 T=5000; GM= .75 Region key T=3000; GM= .73 T=6000; GM= .08

Done so far Used the metric on a block-wise walevetbased segmentation algorithm on some sample mosaic To be done Fully test the metric on a wide range of segmentation algorithms Decouple the various components of the metric and test the individual performance measures instead of the overall score Extend the metric to measure one region vs background segmentation

Recently Viewed Presentations

  • Training Agenda/Objectives  Day 1: Cost Management Overview  Understanding

    Training Agenda/Objectives Day 1: Cost Management Overview Understanding

    Confirmations within GFEBS are a form of Direct Activity Allocations. Used to reflect the DPW work to an order for Equipment, GSA Vehicles, and DPW work hrs (DPW Hrs to work orders are not charged out in ATAAPS) ... Master...
  • A Simple Grammar - IIT Bombay

    A Simple Grammar - IIT Bombay

    The English labiodental fricatives [f] and [v] are produced by pressing the lower lip against the upper teeth, allowing a restricted airflow between the upper teeth. The dental fricatives [th] and [dh] allow air to flow around the tongue between...
  • Donald Bren School of Information and Computer Sciences

    Donald Bren School of Information and Computer Sciences

    Only School of Computing in the UC * UCIRVINE. UCIrvine. The Bren School of ICS is the only computing-focused school in the University of California's 10-campus system.. The School's independence (as opposed to being a department within an engineering school)...
  • Exam Success How to help your child with

    Exam Success How to help your child with

    Making a flow chart. This technique is helpful when you need to remember events in order, like a process in science or geography or a story in English: Make flash cards, little cards with a question or prompt on one...
  • Irregular Verbs

    Irregular Verbs

    Past Participle Past Present …. feed fed fed feeds Past Participle Past Present …. feed ? ? ? Past Participle Past Present …. feel felt felt feels Past Participle Past Present …. feel ? ? ? Past Participle Past Present...
  • TWENTY-SECOND SUNDAY IN ORDINARY TIME SEPTEMBER 1, 2013

    TWENTY-SECOND SUNDAY IN ORDINARY TIME SEPTEMBER 1, 2013

    MARY'S SHELTER: Mary's Shelter was founded as a viable alternative to abortion by women and men who believe that every living person, including the unborn, are children of one Heavenly Father who creates all in his infinite goodness and wisdom....
  • 3.2 Empires of India p.76-81 Big Idea: New

    3.2 Empires of India p.76-81 Big Idea: New

    like Mahabharata(muh HAH BAH ruhtuh), longest poem in any written language, 90,000 stanzas, and Ramayana. Both tell about deeds of great warriors. Most famous section was Bhagaved ( Bah guhVahd geetuhGita (sermon by the god Krishna on eve of major...
  • Solubility - Mr. Hoover's Science Classes

    Solubility - Mr. Hoover's Science Classes

    Solubility Categories (P 323) High Solubility - with a maximum concentration at SATP (standard ambient temperature and pressure) of greater than or equal to 0.1 mol/L . Low Solubility - with a maximum concentration of SATP of less than 0.1...