- Email: [email protected]
An Intelligent Generalized System for Tissue Classification By Incorporating Qualitative Medical Knowledge
Proceedings of the 7th IFAC Symposium on Modelling and Control in Biomedical Systems Aalborg, Denmark, August 12-14, 2009
An Intelligent Generalized System for Tissue Classification By Incorporating Qualitative Medical Knowledge H. KANG *, A. PINTI*, A. TALEB-AHMED*, X. ZENG** * Université de Valenciennes, LAMIH-CNRS, Le Mont Houy, 59313 Valenciennes, FRANCE (e-mail: [email protected]). **ENSAIT, 2 allée Louise et Victor Champier, 59100 Roubaix, FRANCE (e-mail: [email protected]). Abstract: In the diagnosis using MRI images, image segmentation techniques play a key role. Existing segmentation methods are generally based on the features such as grey level and texture. However, these methods can’t identify the physical significance of segmented objects from image because the general features such as grey level can not take into consideration the specialized medical knowledge, which is important when doctors study them manually using their own vision and experience. To deal with this problem, many tissue classification systems have been developed by incorporating the specific medical knowledge. All of these systems focus on specific applications and are not normalized and structured. So they lack of certainty and precision when being applied in other contexts. In this paper, we propose an intelligent generalized tissue classification system which combines both the Fuzzy C-Means algorithm and the qualitative medical knowledge on geometric properties of different tissues. In this system, a general geometric model is proposed which permits to formalize non structured and non normalized medical knowledge from various medical images. This system has been successfully applied to the classification of human thigh, crus, arm, forearm, and brain in MRI images. Keywords: Fuzzy C-Means, MRI, Tissue classification, Image segmentation, Medical knowledge. 1. INTRODUCTION Magnetic Resonance Imaging (MRI) allows us to explore the living organs in a non-invasive way. The analysis of MRI images can be used in the quantification of human body composition, such as the quantification of muscle/fat ratio ([Colin et al. 1995]). In the diagnosis using MRI, image segmentation techniques play a key role. It aims at partitioning an image into a number of non-overlapped and constituent regions which are homogeneous with respect to some characteristics such as grey level or texture ([Zhang and Chen 2004]). Many clustering methods such as Fuzzy C-means ([Ahmed et al. 2002, Chen and Zhang 2004, Jianzhong et al. 2008, Li et al. 2008, Zhang and Chen 2004]) have been proposed for image segmentation. However, these segmentation methods can’t identify the physical significance of segmented objects from image. This is because the general features that they used, such as grey level and texture, can not take into consideration the specialized medical knowledge, which is crucial when doctors study them manually using their own vision and experience. Therefore, it is necessary to incorporate our a priori knowledge on medical image analysis in order to interpret the segmented objects or classes. In this context, various knowledge-based tissue classification systems have been proposed, such as the analysis for brain tissues ([Li et al. 1993, Li et al. 1994, Li et al. 1995]), the analysis for bone composition ([Liu et al. 1999]), and the diagnosis for chest ([Brown et al. 1997]). Nevertheless, all of these systems focus on specific applications and are not normalized and structured. So they lack of certainty and
978-3-902661-49-4/09/$20.00 © 2009 IFAC
precision when being applied in other contexts. Therefore, a more generalized automatic tissue classification system is needed for integrating medical knowledge in a flexible way and being adapted to different applications. In this paper, we propose an intelligent generalized tissue classification system which combines both the Fuzzy C-Means algorithm and the qualitative medical knowledge on geometric properties of different tissues for further improving obtained segmentation results. There are three levels in this system. The low lewel is Knowledge Acquisition, including a Knowledge Base module and an Interface module. The user friendly interface is constructed so that medical knowledge can be integrated into this system in an interactive way. The middle level is Rules Generation, it includes the Geometric Models for formalizing medical knowledge, and Rules for splitting each class obtained by Fuzzy C-Means algorithm into tissues and giving significance to each class. The high level is the Rules Control Strategy Generation, we propose two principles to define the priorities for these rules in order to optimize their application. The rules with higher priorities will be applied before those with lower priorities. Using this system, we can effectively label tissues and give significance to segmented classes from image which can be used in further medical analysis. 2. THE PROPOSED SYSTEM There are three levels in this system. The low lewel is Knowledge Acquisition. The middle level is Rules Generation. The high level is the Rules Control Strategy Generation.
266
10.3182/20090812-3-DK-2006.0091
7th IFAC MCBMS (MCBMS'09) Aalborg, Denmark, August 12-14, 2009
2.1 Low level: Knowledge Acquisition The low level includes a Knowledge Base module and an Interface module. The Knowledge Base is composed of a number of features used for classifying each tissue. The knowledge is acquired through three channels. The first is a documented discussion with a doctor which aides in deciding what are the important anatomical features that need to be modeled. The second is the use of anatomical text books and anatomical atlases. And the third is an informal discussion with doctor and radiologists. These channels allow to record any anomalies in the representation of the anatomy of tissues. The Interface module is an online questionnaire. The experts are invited to fill this questionnaire or we search for the answers from the anatomical text books. In this system, the knowledge is expressed using linguistic description. We take human crus as an example to show this questionnaire. For example, (a). Is spongy bone inside cortical bone? (b). Do muscle and adipose tissue border each other? From all the answers collected in the questionnaire, we concentrate on six features, of which the Knowledge Base module is composed: (a). Relative positions between tissues. (b). Neighboring relation of each tissue. (c). Ranking order of all tissues according to their areas. (d). Number of connected components of certain tissues. (e). Tissues having similar grey level. (f). Shape of certain tissue.
LTA=(Ti1, Ti2, …, Tiq) where Area(Tij)>Area(Tik) for j
(d). The set of Numbers of Connected components in tissues. NC={ | i˛ {1, …, n}} where nci is the number of connected components in tissue Ti. (e). The set of Groups of Tissues having similar grey levels. GT={GTi | i˛ {1, …, p} | p is number of groups of grey levels } and Grey_level(Tj)=Grey_level(Tk) for any two tissues Tj, Tk˛ GTi (j≠k). GTi is the ith group of tissues having similar grey levels. (f). The Set of shapes for the Tissues with one connected component ST={ | i˛ {1, …, n} and Si˛ {triangle, circle, rectangle, …}. After having formalized all the six features, we store them in a predefined data structure, i.e. a 2-dimensional array. We now take the feature “The set of relative positions between tissues” as an example to show how we store them. Name of array: RelativePosition. The data structure is a n×3 array in the following form:
2.2 Middle level: Rules Generation
Attribute
There are two parts in middle level, Geometric Models and Rules. In the Geometric model, we formalize the above six features and then store them in a predefined data structure, i.e. a 2-dimensional array. Based on these formalized features, we set up in the Rules module one rule for each feature in order to split each class obtained by Fuzzy C-Means algorithm ([Ahmed et al. 2002]) into tissues and giving significance to each class. In Geometric Models, assuming there exist n tissues on a specific medical image denoted as T1, …, Tn, we formalize the knowledge on geometric properties of tissues as follows. (a). The set of relative positions between tissues. RP={ | i, j˛ {1, …n,} and Ti≠Tj} where A is an attribute between two tissues and A˛ {above, below, left, right, above left, below left, above right, below right, inside, outside}. For example, “” means that Tissue1 is above Tissue2. (b). The matrix of Neighbors for Tissues. It is a n×n matrix denoted as NT={ntij}. i, j˛ {1, …n,}. ntij is the element in the position (i,j) of the matrix. If ntij=1, then Ti and Tj are neighbors. If ntij=0, then Ti and Tj are not neighbors. (c). The List of Tissues with one connected component ranked in the decreasing order of their Areas.
i,
Tissue 1
Tissue 2
Attribute˛ {above, below, left, right, above left, below left, above right, below right, inside, outside}. For example, (above, Tissue 1, Tissue 2) means Tissue 1 is above Tissue 2. When applying the Fuzzy C-Means algorithm for image segmentation, we often obtain several tissues in one segmented class and several classes corresponding to one tissue. This is because the feature used in this algorithm is based on grey levels. A knowledge guided intelligent control procedure is used for splitting obtained classes of pixels denoted as C1, C2, …, Cp into tissues. The predefined number of classes for the Fuzzy C-Means algorithm is p (we define this number by our a priori knowledge on images), i.e. the number of groups each including tissues of similar grey levels. In this case, we obtain classes each having similar grey levels. For splitting each class into tissues and giving significance to each class, we define the following rules according to the previous geometric model. These rules permit to label tissues from classes. (Rule 1). Background extraction. It is easy to extract and label the background from an image because it is an object with only one connected component linking all the four bounds of the image. (Rule 2). Relative positions of tissues. For any labeled tissue Ti, if there is only one unlabeled tissue Tj so that˛ RP (A can be any relation attribute), and if Tj does not have similar grey levels with any other
267
7th IFAC MCBMS (MCBMS'09) Aalborg, Denmark, August 12-14, 2009
tissues, then we can find the class corresponding to Tj by comparing positions of pixels and label it. (Rule 3). Neighbors of tissues. For each unlabeled tissue Ti, if its neighboring tissues have been all labeled, then we can find and label Ti using connectivity analysis related to these neighboring tissues. (Rule 4). The tissues with one connected component ranked in the decreasing order of their areas. If any group of similar grey levels GTi has only one unlabeled tissue, then we can rank all the unlabeled classes with one connected component according to their areas and make equivalence between them and all ranked unlabeled tissues with one connected component. (Rule 5). Numbers of connected components in tissues.
3. VALIDATION OF OUR SYSTEM Our system has been successfully applied to the tissue classification of human arm, forearm, thigh, crus and brain. We take the classification of crus as an example to show this system. Fig.1 shows one MRI image of crus. It’s in axial plane, T1-weighted, obtained from a system of Philips Intera of 1.5 Tesla. The size of voxel is 0.93 * 0.93 * 1.0 mm3, and the coding is 12 bits. In addition, the signal has been analyzed by spin echo technique in order to minimize artifacts. We can see that there are four tissues and one background. And cortical bone and background have similar grey levels, and spongy bone and adipose tissue have similar grey levels. Therefore, using Fuzzy C-Means algorithm with 5 classes results in bad segmentation (Fig. 2).
Given a specific number of connected objects, if the number of unlabeled tissues is 1, and if the number of unlabeled classes is 1, then this class is equivalent to the related tissue.
Cortical bone
(Rule 6). Groups of tissues having similar grey levels.
Spongy bone
For each group of tissues GTi corresponding to one class obtained by the Fuzzy C-Means algorithm, if only one tissue in it has not been labeled, then we can easily separate and label it by deducing the pixels corresponding to labeled tissues from this class.
Muscle Adipose tissue
(Rule 7). Shapes of the tissues with one connected component. For each shape, if the number of unlabeled tissues in ST is 1, and if the number of corresponding unlabeled classes with one connected component is 1, then this class is equivalent to the related tissue.
Background
Fig. 1. One MRI image of human crus.
2.3 High level: Rules Control Strategy Generation When we run our system, we first use the clustering algorithm to obtain image segmentation. Next, we can perform the classification by iteratively applying the above seven rules. In order to optimize the application of these rules, we define three priorities for them. The rules with higher priorities will be applied before those with lower priorities. Two principles are proposed to define the priorities. (a). The priority of a rule can be defined according to the number of the corresponding premises. Rules with fewer number of premises have higher priority than those with bigger number of premises. Then, we have: (1). Rule 1 has the highest priority, (2). Rules 2, 5, and 7 have the lowest priority, and (3). the priority of Rules 3, 4, 6 is medium. (b). The priority of a rule can be defined according to the number of arrays involved. Rules with fewer number of arrays involved have higher priority than those with bigger number of arrays involved. So we have: (1). Rule 1 has the highest priority, (2). Rules 2, 3, and 4 have the lowest priority, and (3). the priority of Rules 5, 6, and 7 is medium.
Fig. 2. Segmentation by Fuzzy C-Means with five classes. When applying the proposed system, we run the Fuzzy C-Means algorithm with 3 classes because only three grey levels can be observed (p=3). Then, we obtain from the clustering result that cortical bone and background belong to one class C1, adipose tissue and spongy bone to another class C2, and muscle corresponds to C3. We use the first principle to define the priorities of rules. So we first use Rule 1 which is with the highest priority to separate the background from C1. Then we recursively apply the Rule 3, Rule 4 and Rule 6 which are with medium priority for separating new tissues from the obtained classes. We use Rule 6 (groups of tissues having similar grey levels) to label cortical bone from the remaining pixels in C1 with respect to the background. Next, we apply the Rule 3 (neighbors of tissues) with respect to the cortical bone for determining pixels of spongy bone from C2. The remaining pixels of C2 can be then identified as adipose
268
7th IFAC MCBMS (MCBMS'09) Aalborg, Denmark, August 12-14, 2009
tissue using Rule 6 with respect to the spongy bone. Using Rule 3 with respect to cortical bone and adipose tissue, we can label C3 as muscle. Using our proposed intelligent system, we effectively separate the four tissues and the background on the image of crus. Then we reconstruct a synthetic image using the obtained five classes with following grey levels: background: 20; adipose tissue: 250; spongy bone: 150; cortical bone: 200; muscle: 60 (see Fig. 3a). The quality in Fig. 3a is much higher than the one in Fig. 2.
legs based on the obtained results. We will also validate our system on other parts of human body. Furthermore, we will make more efforts for exploring other common tissue features and add new rules to our system so that our system can be more generalized. We will also endeavor to apply new methods to formalize the tissue geometric feature, such as the Fuzzy Logic. In addition, we will try to propose new principles to define the priorities of rules in order to optimize their applications. ACKNOWLEDGEMENT This work was financed by the Conseil Régional Nord Pas de Calais, France. REFERENCES
(a)
(b)
Fig. 3. Classification result of crus. (a). The classification by our system. (b). The classification by a medical expert. Next, we do quantitative comparison of the classification by our system and a reference classification (see Fig.3b) provided by a medical expert in Lille Regional University Hospital (CHRU de Lille, France). The comparison criterion (cc) is: cc =
A ij ˙ A refj A ij ¨ A refj
(1)
where Aij is the set of pixels belonging to the jth class obtained by the ith method, and Arefj is the set of pixels belonging to jth class in the reference image. Table 1 shows the corresponding quantitative comparison. Table 1. Quantitative comparison of classification Adipose tissue 93.18%
Cortical bone 81.96%
Spongy bone
Muscle
88.07%
98.05%
4. CONCLUSION In this paper, we present a more generalized automatic system for tissue classification which can be adapted to different parts of human body. In this system, a general geometric model is proposed which permits to formalize non structured and non normalized medical knowledge from various medical images. An intelligent control procedure is developed for transforming the medical knowledge into several rules in order to improve the quality of the segmentation by Fuzzy C-Means. Two principles are proposed to define the priorities for these rules in order to optimize their application. In future, we will work on the quantification of the muscle/fat ratio, assessment of the muscle/fat temporal variation, and measurement of the volume of muscle, fat and bone in human
Ahmed, M.N., Yamany, S.M., Mohamed, N., Farag, A.A., and Moriarty, T. (2002). “A modified fuzzy c-means algorithm for bias field estimation and segmentation of MRI data”. IEEE TMI, Vol.21, N° 3, pp.193-200. Brown, M.S., McNitt-Gray, M.F., Mankovich, N.J., Goldin, J.G., Hiller, J., Wilson, L.S., and Aberle, D.R. (1997). “Method for segmenting chest CT image data using an anatomical model: Preliminary results”. IEEE TMI, Vol.16, N° 6, pp.828-839. Chen, S., and Zhang, D. (2004). “Robust image segmentation using FCM with spatial constraints based on new kernel-induced distance measure”. IEEE TSMC-Part B: Cybernetics, Vol.34, N° 4, pp.1907-1916. Colin, A., Erbland, E., Datin, C., Boire, J.Y., Veyre, A., and Zanca, M. (1995). “Automatic muscle/fat quantification on MR images”. 17th International Conference of the IEEE EMB, Montreal, Vol.1, pp.479-480. Jianzhong, W., Jun, K., Yinghua, L., Miao, Q., and Baoxue, Z. (2008). “A modified FCM algorithm for MRI brain image segmentation using both local and non-local spatial constraints”. CMIG, Vol.32, pp.685-698. Li, C., Goldgof, D.B., and Hall, L.O. (1993). “Knowledge based classification and tissue labeling of MR images of human brain”. IEEE TMI, Vol.12, N° 4, pp.740-750. Li, H., Deklerck, R., Cuyper, B.D., Hermanus, A., Nysser, E., and Cornelis, J. (1995). “Object recognition in brain CT-scans: knowledge-based fusion of data from multiple feature extractors”. IEEE TMI, Vol.14, N° 2, pp.212-229. Li, H., Deklerck, R., and Cornelis, J. (1994). “Integration of multiple knowledge sources in a system for brain CT-scan interpretation based on the blackboard model”. 10th Conference on AIA, San Antonio, pp.336- 343. Li, X., Zhang, T., and Qu, Z.(2008). “Image segmentation using fuzzy clustering with spatial constraints based on markov random field via bayesian theory”. Signal Processing for Audio and Visual Systems and its Implementations, Vol.E91-A, N° 3, pp.723-728. Liu, Z.Q., Liew, H.L., Clement, J.G., and Thomas, C.D.L. (1999). “Bone image segmentation”. IEEE TBE, Vol.46, N° 5, pp.565-570. Zhang, D., and Chen, S. (2004). “A novel kernelized fuzzy C-means algorithm with application in medical image segmentation”. AIM, Vol.32, pp.37-50.
269
An Intelligent Generalized System for Tissue Classification By Incorporating Qualitative Medical Knowledge H. KANG *, A. PINTI*, A. TALEB-AHMED*, X. ZENG** * Université de Valenciennes, LAMIH-CNRS, Le Mont Houy, 59313 Valenciennes, FRANCE (e-mail: [email protected]). **ENSAIT, 2 allée Louise et Victor Champier, 59100 Roubaix, FRANCE (e-mail: [email protected]). Abstract: In the diagnosis using MRI images, image segmentation techniques play a key role. Existing segmentation methods are generally based on the features such as grey level and texture. However, these methods can’t identify the physical significance of segmented objects from image because the general features such as grey level can not take into consideration the specialized medical knowledge, which is important when doctors study them manually using their own vision and experience. To deal with this problem, many tissue classification systems have been developed by incorporating the specific medical knowledge. All of these systems focus on specific applications and are not normalized and structured. So they lack of certainty and precision when being applied in other contexts. In this paper, we propose an intelligent generalized tissue classification system which combines both the Fuzzy C-Means algorithm and the qualitative medical knowledge on geometric properties of different tissues. In this system, a general geometric model is proposed which permits to formalize non structured and non normalized medical knowledge from various medical images. This system has been successfully applied to the classification of human thigh, crus, arm, forearm, and brain in MRI images. Keywords: Fuzzy C-Means, MRI, Tissue classification, Image segmentation, Medical knowledge. 1. INTRODUCTION Magnetic Resonance Imaging (MRI) allows us to explore the living organs in a non-invasive way. The analysis of MRI images can be used in the quantification of human body composition, such as the quantification of muscle/fat ratio ([Colin et al. 1995]). In the diagnosis using MRI, image segmentation techniques play a key role. It aims at partitioning an image into a number of non-overlapped and constituent regions which are homogeneous with respect to some characteristics such as grey level or texture ([Zhang and Chen 2004]). Many clustering methods such as Fuzzy C-means ([Ahmed et al. 2002, Chen and Zhang 2004, Jianzhong et al. 2008, Li et al. 2008, Zhang and Chen 2004]) have been proposed for image segmentation. However, these segmentation methods can’t identify the physical significance of segmented objects from image. This is because the general features that they used, such as grey level and texture, can not take into consideration the specialized medical knowledge, which is crucial when doctors study them manually using their own vision and experience. Therefore, it is necessary to incorporate our a priori knowledge on medical image analysis in order to interpret the segmented objects or classes. In this context, various knowledge-based tissue classification systems have been proposed, such as the analysis for brain tissues ([Li et al. 1993, Li et al. 1994, Li et al. 1995]), the analysis for bone composition ([Liu et al. 1999]), and the diagnosis for chest ([Brown et al. 1997]). Nevertheless, all of these systems focus on specific applications and are not normalized and structured. So they lack of certainty and
978-3-902661-49-4/09/$20.00 © 2009 IFAC
precision when being applied in other contexts. Therefore, a more generalized automatic tissue classification system is needed for integrating medical knowledge in a flexible way and being adapted to different applications. In this paper, we propose an intelligent generalized tissue classification system which combines both the Fuzzy C-Means algorithm and the qualitative medical knowledge on geometric properties of different tissues for further improving obtained segmentation results. There are three levels in this system. The low lewel is Knowledge Acquisition, including a Knowledge Base module and an Interface module. The user friendly interface is constructed so that medical knowledge can be integrated into this system in an interactive way. The middle level is Rules Generation, it includes the Geometric Models for formalizing medical knowledge, and Rules for splitting each class obtained by Fuzzy C-Means algorithm into tissues and giving significance to each class. The high level is the Rules Control Strategy Generation, we propose two principles to define the priorities for these rules in order to optimize their application. The rules with higher priorities will be applied before those with lower priorities. Using this system, we can effectively label tissues and give significance to segmented classes from image which can be used in further medical analysis. 2. THE PROPOSED SYSTEM There are three levels in this system. The low lewel is Knowledge Acquisition. The middle level is Rules Generation. The high level is the Rules Control Strategy Generation.
266
10.3182/20090812-3-DK-2006.0091
7th IFAC MCBMS (MCBMS'09) Aalborg, Denmark, August 12-14, 2009
2.1 Low level: Knowledge Acquisition The low level includes a Knowledge Base module and an Interface module. The Knowledge Base is composed of a number of features used for classifying each tissue. The knowledge is acquired through three channels. The first is a documented discussion with a doctor which aides in deciding what are the important anatomical features that need to be modeled. The second is the use of anatomical text books and anatomical atlases. And the third is an informal discussion with doctor and radiologists. These channels allow to record any anomalies in the representation of the anatomy of tissues. The Interface module is an online questionnaire. The experts are invited to fill this questionnaire or we search for the answers from the anatomical text books. In this system, the knowledge is expressed using linguistic description. We take human crus as an example to show this questionnaire. For example, (a). Is spongy bone inside cortical bone? (b). Do muscle and adipose tissue border each other? From all the answers collected in the questionnaire, we concentrate on six features, of which the Knowledge Base module is composed: (a). Relative positions between tissues. (b). Neighboring relation of each tissue. (c). Ranking order of all tissues according to their areas. (d). Number of connected components of certain tissues. (e). Tissues having similar grey level. (f). Shape of certain tissue.
LTA=(Ti1, Ti2, …, Tiq) where Area(Tij)>Area(Tik) for j
(d). The set of Numbers of Connected components in tissues. NC={
2.2 Middle level: Rules Generation
Attribute
There are two parts in middle level, Geometric Models and Rules. In the Geometric model, we formalize the above six features and then store them in a predefined data structure, i.e. a 2-dimensional array. Based on these formalized features, we set up in the Rules module one rule for each feature in order to split each class obtained by Fuzzy C-Means algorithm ([Ahmed et al. 2002]) into tissues and giving significance to each class. In Geometric Models, assuming there exist n tissues on a specific medical image denoted as T1, …, Tn, we formalize the knowledge on geometric properties of tissues as follows. (a). The set of relative positions between tissues. RP={
i,
Tissue 1
Tissue 2
Attribute˛ {above, below, left, right, above left, below left, above right, below right, inside, outside}. For example, (above, Tissue 1, Tissue 2) means Tissue 1 is above Tissue 2. When applying the Fuzzy C-Means algorithm for image segmentation, we often obtain several tissues in one segmented class and several classes corresponding to one tissue. This is because the feature used in this algorithm is based on grey levels. A knowledge guided intelligent control procedure is used for splitting obtained classes of pixels denoted as C1, C2, …, Cp into tissues. The predefined number of classes for the Fuzzy C-Means algorithm is p (we define this number by our a priori knowledge on images), i.e. the number of groups each including tissues of similar grey levels. In this case, we obtain classes each having similar grey levels. For splitting each class into tissues and giving significance to each class, we define the following rules according to the previous geometric model. These rules permit to label tissues from classes. (Rule 1). Background extraction. It is easy to extract and label the background from an image because it is an object with only one connected component linking all the four bounds of the image. (Rule 2). Relative positions of tissues. For any labeled tissue Ti, if there is only one unlabeled tissue Tj so that
267
7th IFAC MCBMS (MCBMS'09) Aalborg, Denmark, August 12-14, 2009
tissues, then we can find the class corresponding to Tj by comparing positions of pixels and label it. (Rule 3). Neighbors of tissues. For each unlabeled tissue Ti, if its neighboring tissues have been all labeled, then we can find and label Ti using connectivity analysis related to these neighboring tissues. (Rule 4). The tissues with one connected component ranked in the decreasing order of their areas. If any group of similar grey levels GTi has only one unlabeled tissue, then we can rank all the unlabeled classes with one connected component according to their areas and make equivalence between them and all ranked unlabeled tissues with one connected component. (Rule 5). Numbers of connected components in tissues.
3. VALIDATION OF OUR SYSTEM Our system has been successfully applied to the tissue classification of human arm, forearm, thigh, crus and brain. We take the classification of crus as an example to show this system. Fig.1 shows one MRI image of crus. It’s in axial plane, T1-weighted, obtained from a system of Philips Intera of 1.5 Tesla. The size of voxel is 0.93 * 0.93 * 1.0 mm3, and the coding is 12 bits. In addition, the signal has been analyzed by spin echo technique in order to minimize artifacts. We can see that there are four tissues and one background. And cortical bone and background have similar grey levels, and spongy bone and adipose tissue have similar grey levels. Therefore, using Fuzzy C-Means algorithm with 5 classes results in bad segmentation (Fig. 2).
Given a specific number of connected objects, if the number of unlabeled tissues is 1, and if the number of unlabeled classes is 1, then this class is equivalent to the related tissue.
Cortical bone
(Rule 6). Groups of tissues having similar grey levels.
Spongy bone
For each group of tissues GTi corresponding to one class obtained by the Fuzzy C-Means algorithm, if only one tissue in it has not been labeled, then we can easily separate and label it by deducing the pixels corresponding to labeled tissues from this class.
Muscle Adipose tissue
(Rule 7). Shapes of the tissues with one connected component. For each shape, if the number of unlabeled tissues in ST is 1, and if the number of corresponding unlabeled classes with one connected component is 1, then this class is equivalent to the related tissue.
Background
Fig. 1. One MRI image of human crus.
2.3 High level: Rules Control Strategy Generation When we run our system, we first use the clustering algorithm to obtain image segmentation. Next, we can perform the classification by iteratively applying the above seven rules. In order to optimize the application of these rules, we define three priorities for them. The rules with higher priorities will be applied before those with lower priorities. Two principles are proposed to define the priorities. (a). The priority of a rule can be defined according to the number of the corresponding premises. Rules with fewer number of premises have higher priority than those with bigger number of premises. Then, we have: (1). Rule 1 has the highest priority, (2). Rules 2, 5, and 7 have the lowest priority, and (3). the priority of Rules 3, 4, 6 is medium. (b). The priority of a rule can be defined according to the number of arrays involved. Rules with fewer number of arrays involved have higher priority than those with bigger number of arrays involved. So we have: (1). Rule 1 has the highest priority, (2). Rules 2, 3, and 4 have the lowest priority, and (3). the priority of Rules 5, 6, and 7 is medium.
Fig. 2. Segmentation by Fuzzy C-Means with five classes. When applying the proposed system, we run the Fuzzy C-Means algorithm with 3 classes because only three grey levels can be observed (p=3). Then, we obtain from the clustering result that cortical bone and background belong to one class C1, adipose tissue and spongy bone to another class C2, and muscle corresponds to C3. We use the first principle to define the priorities of rules. So we first use Rule 1 which is with the highest priority to separate the background from C1. Then we recursively apply the Rule 3, Rule 4 and Rule 6 which are with medium priority for separating new tissues from the obtained classes. We use Rule 6 (groups of tissues having similar grey levels) to label cortical bone from the remaining pixels in C1 with respect to the background. Next, we apply the Rule 3 (neighbors of tissues) with respect to the cortical bone for determining pixels of spongy bone from C2. The remaining pixels of C2 can be then identified as adipose
268
7th IFAC MCBMS (MCBMS'09) Aalborg, Denmark, August 12-14, 2009
tissue using Rule 6 with respect to the spongy bone. Using Rule 3 with respect to cortical bone and adipose tissue, we can label C3 as muscle. Using our proposed intelligent system, we effectively separate the four tissues and the background on the image of crus. Then we reconstruct a synthetic image using the obtained five classes with following grey levels: background: 20; adipose tissue: 250; spongy bone: 150; cortical bone: 200; muscle: 60 (see Fig. 3a). The quality in Fig. 3a is much higher than the one in Fig. 2.
legs based on the obtained results. We will also validate our system on other parts of human body. Furthermore, we will make more efforts for exploring other common tissue features and add new rules to our system so that our system can be more generalized. We will also endeavor to apply new methods to formalize the tissue geometric feature, such as the Fuzzy Logic. In addition, we will try to propose new principles to define the priorities of rules in order to optimize their applications. ACKNOWLEDGEMENT This work was financed by the Conseil Régional Nord Pas de Calais, France. REFERENCES
(a)
(b)
Fig. 3. Classification result of crus. (a). The classification by our system. (b). The classification by a medical expert. Next, we do quantitative comparison of the classification by our system and a reference classification (see Fig.3b) provided by a medical expert in Lille Regional University Hospital (CHRU de Lille, France). The comparison criterion (cc) is: cc =
A ij ˙ A refj A ij ¨ A refj
(1)
where Aij is the set of pixels belonging to the jth class obtained by the ith method, and Arefj is the set of pixels belonging to jth class in the reference image. Table 1 shows the corresponding quantitative comparison. Table 1. Quantitative comparison of classification Adipose tissue 93.18%
Cortical bone 81.96%
Spongy bone
Muscle
88.07%
98.05%
4. CONCLUSION In this paper, we present a more generalized automatic system for tissue classification which can be adapted to different parts of human body. In this system, a general geometric model is proposed which permits to formalize non structured and non normalized medical knowledge from various medical images. An intelligent control procedure is developed for transforming the medical knowledge into several rules in order to improve the quality of the segmentation by Fuzzy C-Means. Two principles are proposed to define the priorities for these rules in order to optimize their application. In future, we will work on the quantification of the muscle/fat ratio, assessment of the muscle/fat temporal variation, and measurement of the volume of muscle, fat and bone in human
Ahmed, M.N., Yamany, S.M., Mohamed, N., Farag, A.A., and Moriarty, T. (2002). “A modified fuzzy c-means algorithm for bias field estimation and segmentation of MRI data”. IEEE TMI, Vol.21, N° 3, pp.193-200. Brown, M.S., McNitt-Gray, M.F., Mankovich, N.J., Goldin, J.G., Hiller, J., Wilson, L.S., and Aberle, D.R. (1997). “Method for segmenting chest CT image data using an anatomical model: Preliminary results”. IEEE TMI, Vol.16, N° 6, pp.828-839. Chen, S., and Zhang, D. (2004). “Robust image segmentation using FCM with spatial constraints based on new kernel-induced distance measure”. IEEE TSMC-Part B: Cybernetics, Vol.34, N° 4, pp.1907-1916. Colin, A., Erbland, E., Datin, C., Boire, J.Y., Veyre, A., and Zanca, M. (1995). “Automatic muscle/fat quantification on MR images”. 17th International Conference of the IEEE EMB, Montreal, Vol.1, pp.479-480. Jianzhong, W., Jun, K., Yinghua, L., Miao, Q., and Baoxue, Z. (2008). “A modified FCM algorithm for MRI brain image segmentation using both local and non-local spatial constraints”. CMIG, Vol.32, pp.685-698. Li, C., Goldgof, D.B., and Hall, L.O. (1993). “Knowledge based classification and tissue labeling of MR images of human brain”. IEEE TMI, Vol.12, N° 4, pp.740-750. Li, H., Deklerck, R., Cuyper, B.D., Hermanus, A., Nysser, E., and Cornelis, J. (1995). “Object recognition in brain CT-scans: knowledge-based fusion of data from multiple feature extractors”. IEEE TMI, Vol.14, N° 2, pp.212-229. Li, H., Deklerck, R., and Cornelis, J. (1994). “Integration of multiple knowledge sources in a system for brain CT-scan interpretation based on the blackboard model”. 10th Conference on AIA, San Antonio, pp.336- 343. Li, X., Zhang, T., and Qu, Z.(2008). “Image segmentation using fuzzy clustering with spatial constraints based on markov random field via bayesian theory”. Signal Processing for Audio and Visual Systems and its Implementations, Vol.E91-A, N° 3, pp.723-728. Liu, Z.Q., Liew, H.L., Clement, J.G., and Thomas, C.D.L. (1999). “Bone image segmentation”. IEEE TBE, Vol.46, N° 5, pp.565-570. Zhang, D., and Chen, S. (2004). “A novel kernelized fuzzy C-means algorithm with application in medical image segmentation”. AIM, Vol.32, pp.37-50.
269