Medical Image Segmentation Using Hidden Markov Random Field A Distributed Approach
Download as PPTX, PDF1 like860 views
The document discusses a method for medical image segmentation using Hidden Markov Random Fields (HMRF) presented at the ICDIPC 2013 conference. It outlines the challenges of processing 3D medical images and demonstrates how the implemented method improves segmentation quality and processing time compared to traditional techniques. Experimental results indicate that using a cluster of PCs enhances performance, and future work suggests further optimization and comparison with alternative methods.
Medical Image Segmentation Using Hidden Markov Random Field A Distributed Approach
- 1. The Third International Conference on Digital Information Processing and Communications (ICDIPC 2013) Medical Image Segmentation Using Hidden Markov Random Field A Distributed Approach Theme
- 2. INTRODUCTION SEGMENTATION BY USING HMRF EXPERIMENTAL RESULTS CONCLUSION AND PERSPECTIVES 2
- 3. One exam by CT (Computed Tomography) scanner can produce hundred images. All of these images represents a 3D image Processing and analysis of these images becomes a difficult and daunting task The classical analysis of medical cuts 3 I N T R O D U C T I O N Problem
- 4. 3D automatic segmentation The 3D image The segmented 3D image 4 I N T R O D U C T I O N Solution : Tool to aid the physician to make the decision based on Automatic segmentation.
- 5. 5 T H E A I M Relevance of the physician aid tool to make the decision based on The time of computation The quality of segmentation TIME + QUALITY
- 6. INTRODUCTION SEGMENTATION BY USING HMRF EXPERIMENTAL RESULTS CONCLUSION AND PERSPECTIVES 6
- 7. S E G M E N TAT I O N B Y U S I N G H M R F 7 1 2 3 4 Y: Observed Image X: Hidden Image 2C,s 2 s ),(2-(1)2ln( 2 )²-(y y)(x, t tsx Ss x x xx Ts s s y)(x,minarg Xx x
- 8. S E G M E N TAT I O N B Y U S I N G H M R F 8 Optimizations techniques are used like ICM, … Problem Minimizing the function (x,y) is computationally intractable. Solution
- 9. S E G M E N TAT I O N B Y U S I N G H M R F ICM Algorithm: 1. Initialization: Start with an arbitrary labeling x0 and let n=0. 2. At step n: Visit all the sites according to a visiting scheme and in every site : , 3. Increment n. Goto 2, until a stopping criterion is satisfied. 9 ( ) 1 arg min ( )card S n s s s x x U x
- 10. INTRODUCTION SEGMENTATION BY USING HMRF EXPERIMENTAL RESULTS CONCLUSION AND PERSPECTIVES 10
- 11. E X P E R I M E N TA L R E S U LT S 11 Configuration Hardware : The cluster of eight identical machines Switch (Catalyst 3560G) Configuration Software: The Parallelization library is Open MPI Platform application framework Qt Linux system (ubuntu 11.04)
- 12. E X P E R I M E N TA L R E S U LT S 12 Benchmark Name of benchmark Dimension Link 1 MRI Phantom 8Bits (t1_icbm_normal_1mm_pn0 _rf0.rawb) 181 x 217 x 181 http://mouldy.bic.mni. mcgill.ca/brainweb/anat omic_normal.html 2 Head MRT Angiography 8Bits (mrt8_angio2.raw) 256 x 320 x 128 http://www.gris.uni- tuebingen.de/edu/areas/s civis/volren/datasets/ne w.html 3 Head MRI CISS 8Bits (mri_ventricles.raw) 256 x 256 x 124 http://www.gris.uni- tuebingen.de/edu/areas/s civis/volren/datasets/ne w.html Benchmarks images used in our tests.
- 13. E X P E R I M E N TA L R E S U LT S 13 Visual results Benchmark : 1
- 14. E X P E R I M E N TA L R E S U LT S 14 Visual results Benchmark : 2
- 15. E X P E R I M E N TA L R E S U LT S 15 Visual results Benchmark : 3
- 16. Evaluating the quality of the segmentation 16 FNFPTP2 TP2DC Kappa index Ground truth The image to segment The segmented image E X P E R I M E N TA L R E S U LT S
- 17. E X P E R I M E N TA L R E S U LT S Comparison : Mean kappa index values Benchmark : 1 Slices : 90-119 Methods : Otsu, MoG, MoGG and our method 17 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 White Matter Gray Matter CSF Matter Otsu MoG MoGG Our Method Methods Kappa Index
- 18. Speed-up 18 E X P E R I M E N TA L R E S U LT S Processing Time
- 19. 19 E X P E R I M E N TA L R E S U LT S Processing Time 0 1 2 3 4 5 6 7 8 9 1 PC 2 PCs 4 PCs 8 PCs Benchmark 1 Benchmark 2 Benchmark 3 Time (h) Number of PCs Benchmarks
- 20. 20 E X P E R I M E N TA L R E S U LT S SPEED UP 0 1 2 3 4 5 6 7 8 9 1 PC 2 PCs 4 PCs 8 PCs Benchmark 1 Benchmark 2 Benchmark 3 Speed-up Number of PCs Benchmarks
- 21. INTRODUCTION SEGMENTATION BY USING HMRF EXPERIMENTAL RESULTS CONCLUSION AND PERSPECTIVES 21
- 22. The kappa index can be used only when we know beforehand segmentation ground truth . In our tests we notice our implemented method seems generally better than the thresholding-based segmentation methods (Otsu, MoG, MoGG ). The processing time is improved by the use of a cluster of PCs. 22 C O N C L U S I O N A N D P E R S P E C T I V E S
- 23. However, further work must take into account like : The cluster of PCs must be incremented to see the limits of its contribution. Comparison with other methods Implementation of other optimization methods 23 C O N C L U S I O N A N D P E R S P E C T I V E S
- 24. Thank you for your attention
Editor's Notes
- #2: Hello everybody, I will present my work titled : Medical Image Segmentation Using Hidden Markov Random Field A Distributed Approach
- #3: and for that we are following the plan as follows : We begin by INTRODUCTION followed by SEGMENTATION BY USING HMRF and EXPERIMENTAL RESULTS and we finish by CONCLUSION AND PERSPECTIVES
- #4: One exam by CT (Computed Tomography) scanner can produce hundred images Processing and analysis of these images becomes a difficult and daunting task
- #5: So the solution is a Tool to aid the physician to make the decision based on Automatic segmentation.
- #6: Pertinence of the aid tool based on two axis, the time and quality in our research we are interesting by improving the time by usage cluster of PCs and quality by usage HMRF.
- #7: and for that we are following the plan as follows : We begin by INTRODUCTION followed by SEGMENTATION BY USING HMRF and EXPERIMENTAL RESULTS and we finish by CONCLUSION AND PERSPECTIVES
- #8: HMRF is a strong model for image segmentation is to see the image to segment as a realization of a Markov Random Field Y={Ys}sS. And The segmented image is seen as the realization of another Markov Random Field X , can be found it by maximizing the function (x,y).
- #9: Minimizing the function (x,y) is computationally intractable, so we need some optimization techniques.
- #10: For example ICM method we can summarize it by three great line the first step is initialization and the second step is looking for the minimum locally until a stopping criterion is satisfied.
- #11: and for that we are following the plan as follows : We begin by INTRODUCTION followed by SEGMENTATION BY USING HMRF and EXPERIMENTAL RESULTS and we finish by CONCLUSION AND PERSPECTIVES
- #12: Before we give our EXPERIMENTAL RESULTS It is necessary to give our Configuration Hardware and Software So we have used The cluster of eight identical machines ubuntu like system and open MPI like parallelization library and QT as Platform application framework related by Switch type Catalyst
- #13: Here we show you some benchmarks images we have used in our tests and their dimension.
- #14: Here we show you some visual results of benchmark 1
- #15: And here some visual results of benchmark 2
- #16: And here some visual results of benchmark 3 Our visual results always seems good but we can’t based on visual results to know the good methods for that and to be certain of it we use kappa Index.
- #17: Kappa index Evaluating the quality of the segmentation can be only use it when the ground truth is known a priori. So we compare the ground truth with our segmentation as shown in figure.
- #18: To know if our results are good or not must be compare it with others. Our results seem generally better than some thresholding methods results.
- #19: Speed up is time factor used to know gain of time when we use p processor compared with one processor
- #20: Here we notice that we always earn nearly half the time when we duplicate the number of PCs
- #21: Here we notice the speed up generally is duplicate when the number of PCs is duplicate
- #22: and for that we are following the plan as follows : We begin by INTRODUCTION followed by SEGMENTATION BY USING HMRF and EXPERIMENTAL RESULTS and we finish by CONCLUSION AND PERSPECTIVES