Application of image processing techniques to tissue texture .ppt

1、Application of image processing techniques to tissue texture analysis and image compression,Advisor : Dr. Albert Chi-Shing CHUNG,Presented by Group ACH1 (LAW Wai Kong and LAI Tsz Chung),Computer Science Final Year Project 2004,Overview,Introduction Motivation Objectives Results Classification algori

2、thms: Feature extraction & Classifier selection Software implementation: Conclusion Future Extension Question and Answer Session,Introduction,- Motivation,Diagnosis of cirrhosis:,1) Manual diagnosis of ultrasonic liver image,2) Histological analysis,Invasive,Inaccurate Results dependent on experienc

3、e of sonographers,Both are time consuming,How about computer aided diagnosis system?,In what extent this system assist doctor?,- Objectives,Designated user interface with support of ultrasonic image compression,No pre-image processing is needed,Reduce storage space,Facilitate the diagnosis process,M

4、ulti-severity level classification,Cirrhosis treatment require severity information.,Machine independence,Compatible with different ultrasound scanning machine,Challenge ! How to classify patients?,2 steps,Step 1: Feature Extraction,Firstly, extract useful features from image.,We have examined sever

5、al feature extraction approaches for performance comparison,The most accurate approach will be implemented in our system,Direct comparison of wavelet coefficient (Haar, Symlets, Daubechies),Histogram of wavelet coefficient (Haar, Symlets, Daubechies),Statistic with “Difference on Gaussians” filter,D

6、irect comparison between multi-scale co-occurrence matrix,Statistic with multi-scale approach and co-occurrence matrix,Step 1: Feature Extraction,The six features:,1) The mean gray level,- Inversely proportion to cirrhosis severity. - Affected by the area of normal tumor,2) The first percentile of t

7、he gray level distribution P,First order statistic,- Inversely proportion to cirrhosis severity. - Affected by the present of normal tumor,Co-occurrence matrix statistic,3) Entropy:,4) Contrast:,5) Angular Second Moment:,6) Correlation,6) Morphological based method,Segment out tumor structure from l

8、iver Count the number and circumference of tumor,Input features: normalized to range between 0,1 Category: normalized to range between 0,1 Classification: by setting thresholds base on # category. 1st layer: 5 hyperbolic tangent sigmoid transfer units 2nd layer: 1 linear transfer unit Train function

9、: Levenberg-Marquardt back-propagation Performance: MSE Stopping threshold: 0.01 Maximum training cycle = 200,Step 2: Classifier,Basic requirements: Continuous learning Multi class classification (severity category) Robust Database can update per patient (one pattern).,Secondly, classify patients ba

10、sed on extracted features,3 classifiers were examined,1) k-Nearest Neighbor Classifier,Use the category of k-nearest neighbor in database to classify a new entry.,The features are normalized by standard score.,Distance-weighted.,Choice of distance: SSD / KLD,Physically, KLD measures relative entropy

11、 between PDF,2) Feed-forward Neural Network,A direct continuation of the work on Bayes classifiers, which relies on Parzen windows classifiers.,Setting:,3) Probabilistic Neural Network,It learns to approximate the PDF of the training examples.,The input features are normalized by standard score.,Com

12、monly used in image feature classification,Evaluation of algorithms,Method of evaluating hypothesis: 10-fold cross validation (in MatLab),Problem: Images of the same patient have similar features!,Solution: Use patient ID to partition the data set.,Problem: uneven class distribution in folds!,Soluti

13、on: Partition the patients based on their category, ensure class distribution is similar to original data set.,The features:,Theoretically, morphology is a descriptive feature, but, practically, fine tuning of parameters is needed.,Segmentation parameter (sigma of Gaussian filter, initial marker int

14、ensity) too sensitive to suit all testing cases,Number of tumors was unreasonably fluctuated. (tumors count ranged from 15 to 90),Comparison of best results among all features sets with different classifier:,The data set is captured by Dr. Simon Yu, consultant and adjunct associate professor from De

15、partment of Diagnostic Radiology and Organ Imaging, Prince of Wales Hospital,Evaluation of algorithms,The classifiers:,Accuracy:, all of them have similar results., Depends on features.,Running time (including partition for 732 testing cases):,Pros and Cons,k-NN,Fast Easy to implement,Sensitive to c

16、lass distribution of data set. Size of database is large and linearly increasing.,FFNN,Size of database is a small constant. Robust,Training is slow. ( 40 times of k-NN) Should update per epoch to prevent noise.,PNN,Fast,Highly sensitive to class distribution of data set. Size of database increases

17、linearly.,k-NN,Conclusion,Developed a designated classification system that can contribute to medical aspect Examined different machine independent classification algorithms for multi-severity classification Proposed utilization of multi-resolution statistic with co-occurrence matrix for cirrhosis d

18、etection Realized machine learning and image processing techniques in a real life situation Explored the knowledge about cirrhosis and liver,Future Extension,Clustering of features Fine tuning the parameters of morphological approach Histological findings of cases will be able to improve our system,Question and Answer Session,