- Open Access
Hybrid model for analysis of abnormalities in diabetic cardiomyopathy and diabetic retinopathy related images
© Shaik et al. 2016
- Received: 7 October 2015
- Accepted: 12 April 2016
- Published: 23 April 2016
At present image processing methods hold a noteworthy position in unravelling various medical imaging challenges. The high risk disorders such as diabetic cardiomyopathy and diabetic retinopathy are considered as applications for proposed method. The dictum of this paper is on observing enhancement and segmentation of the cross sectional view of a blood capillary of a right coronary artery image of a diabetic patient and also retinal images. A hybrid model using hybrid morphological reconstruction technique as pre-processing with watershed segmentation method as post-processing is developed in this work.
Diabetic mellitus (DM) is a metabolic disorder that characterized by incapability of the pancreas to control blood glucose concentration. This dilemma results may make out blood glucose levels out of range (Sharifi et al. 2008). Heart ailments are to blame for 80 % of deaths. Nonetheless, there is an increase in recognition that diabetic patients have a medical condition from an additional cardiac insult termed diabetic cardiomyopathy (Hayat and Patel 2004). Diabetic cardiomyopathy is a situation in which an artery wall thickens as the result of an accumulation of fatty materials such as cholesterol.
The leading pathology in diabetic patients is thickening of basement membrane of a blood vessel making blood vessel narrower leading to occlusion of lumen with fatty materials (Kumar and Shaik 2015). Epidemiological and clinical trial data have recognized the larger incidence and prevalence of heart attacks in diabetes even without a noise at all (Asghar et al. 2009) declaring this as a silent killer.
On the other hand diabetic retinopathy deals with the eye problems which results in visual impairment. This disorder persists for a longer period and an utmost care has to be considered to keep diabetes in control to avoid further consequences.
On a technological note image processing is utilized to extract important features from the images, through which better perception of the scene can be obtained for human viewers (Gonzalez and Woods 1992). The biological vision system is one of the most important means of exploration of the world to humans, making complex task easier for betterment of understanding (Peres et al. 2010). There are numerous algorithms that can be utilised for different applications but enhancement and segmentation are considered as most sort out methods for improving the details in an image. It is not possible to judge that any one method is best in Image processing applications but one can use trial and error method as a practical approach for obtaining the perfect results. Image enhancement is a fundamental task in digital image processing and analysis, aiming to improve the appearance of image in terms of human brightness perception (Intajag et al. 2009). Whereas the segmentation is mainly useful in classification of objects and labelling of the features extracted from image for easy analysis. One should look into that processing of images is done without blemishing the integrity of original image.
Due to the imperfection and variations, the appearance of microscopic images is generally not homogeneous. In order to reduce the influence from undesirable variations within, the hybrid morphological reconstruction (HMR) is used to enhance the image. The steps of HMR are described below.
Complement of the image
Complement of the image
The typical objective of this method is based on the concept to find watershed lines. The basic idea is simple, suppose that a hole is punched in each regional minimum and that the entire geography is flooded from below by allowing water rise through the holes at uniform rate (Ravindraiah and Shaik 2010). The entire process is described by a concept that a dam like thing is constructed to avoid merging and flooding may take place when water reaches the top level of dam. Consequently, watershed algorithm extracts the boundaries. In Tang et al. (2011) watershed algorithm was used for segmentation of splats, a collection of pixels with similar color and spatial location.
The images that are obtained from the public image database which are related to diabetes mellitus are considered as the input images. The images chiefly taken into work here are of cross section of the right coronary artery (RCA) and retinal images with anomalies. Generally the images are RGB images. So the RGB images are processed using the MATLAB software and the images undergo several algorithms to get a better output. Initially the RGB image is converted into grey scale to avoid complex calculations. Next step is to perform the gradient magnitude segmentation function. After the above two steps are finished then the main step, watershed transform segmentation is performed. Watershed transform is the region base segmentation method. In this step it fills the gaps present in the images and finally the analyzing the result.
Regional maxima is used to obtain good foreground markers. By applying all these operations we obtained the left area of lumen which is used for the flow of blood. By extending this process, we can alert the common man about the severity and consequences of the disorder if medication is neglected.
Statistical analysis using MIPAV
Medical image processing, analysis, and visualization (MIPAV) which is an open source is utilized to extract the attributes of significance from the images. These attributes are constructive in elucidation of the abnormalities in precise style. The main feature which makes the researcher to use MIPAV is its applicability even on 3D images and quantification aspect. MIPAV is a Java application and can run on any java enabled PCs. This is the product of Center for Information Technology, National Institutes of Health, Bethesda, MD, USA. The performance of the proposed method was rigorously evaluated using quality metrics like area, perimeter, median, standard deviation of intensity, coefficient of skewness.
Decrease in area indicates that ROI i.e., the degree of severity increases.
Decrease in perimeter indicates that ROI i.e., the degree of severity increases.
Increase in the standard deviation indicates that ROI i.e., has been detected with fine edges.
The variation in skewness (decrease or increase) gives the asymmetry of a distribution.
Increase in median shows the average changes of the pixels that occurred in the segmented image.
Quality assessment metrics for input and output images (ROI) under normal, medium and normal conditions
Standard deviation in RGB
Skewness in RGB
Median in RGB
By tabulating and analyzing these parameters normal people can be alerted about the severity of the problem in three stages i.e., normal, medium and severe conditions.
In this study both pre-processing and post processing used belong to the family of Morphological image processing. The experimental results of HMR technique as pre-processing with watershed segmentation method as post-processing are quite suitable for forecasting of narrowing of lumen and retinopathy in diabetic patients. In future development other pre-processing algorithms combined with the implemented post processing method can give perspective results so that the Medical Professionals may make use of this algorithm for earlier detection of the abnormality and this outline in form of a group wise images may be used to alert a common about significance of the problem.
All authors were involved in drafting the article and all authors approved the final version to be submitted for publication. All authors have added an intellectual significant value to the manuscript. FS carried out the collection of Public database of the images and programming for acquiring the results using MATLAB software. AKS and SMA helped in design, interpretation, writing of manuscript and were involved making appropriate corrections wherever required. All authors read and approved the final manuscript.
This work is supported by Sunrise University-Alwar, Rajasthan where the first author is the research scholar and Annamacharya Institute of Technology and Sciences, Rajampet, A.P. for providing research facilities. And the authors are also thankful to Dr. B. Jayabhaskar Rao, Diabetalogist, Nandalur, A.P. for providing the detailed explanation of Diabetes and its abnormalities.
Fahimuddin. Shaik chaired a Session at IEEE International Conference (ICMET-2010) held in Singapore on September 11th 2010 and also in ICCET-12 in 2012 at Kollam, Kerala. He was Scientific Committee member of III Workshop on Technology for Healthcare and Healthy Lifestyle, 2011(WTHS ‘11), Valencia, Spain, 1st–2nd December 2011. He has authored books by titles “Medical Imaging in Diabetes” with Cinnamonteal Publishers in 2011, “Image Processing in Diabetic Related Causes” with Springer Publications in 2015, “Retinal Vessel Classification Via Image Processing Techniques” with VSRD Academic Publishing Ltd., in 2015 and “Signal And Image Processing In Medical Applications” with Springer Publications in 2016. Dr. Anil Kumar Sharma has published technical papers in International journals as well as in international and national conferences. He is Reviewer and Editor of reputed Journals. His research and teaching interest include Microprocessor, VLSI Design, Neuro-Fuzzy Modeling, RADARs and its Data Handling Systems. He has attended 4 days study and Networking Tour at UTP, Malaysia to visit Laboratories and R&D Facilities in the areas of Intelligent Signal and Imaging and Nanotechnology in 2014. He has authored books by titles “Digital Logic Design and Basic Electrical and Electronics Engineering”. Dr. Syed Musthak Ahmed is Vice Chairman, IEEE Education Society Chapter, IEEE Hyderabad Section for the academic year 2015. He was Co-chair for SIESCON11 conference, Dr. M.G.R. Anna University, Chennai. He is a member of ISTE, IEEE, FISSS, and FIETE. He has various publications in National and International Journal/Conferences.
The authors declare that they have no competing interests.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
- Asghar O, Al-Sunni A, Withers S (2009) Diabetic cardiomyopathy. The Manchester Heart Centre, ManchesterGoogle Scholar
- Gonzalez RC, Woods RE (1992) Digital image processing, 1st edn. Addison-Wesley, Reading, MAGoogle Scholar
- Hayat SA, Patel B, Khattar RS, Malik RA (2004) Diabetic cardiomyopathy: mechanisms, diagnosis, and treatment. Clin Sci (Lond) 107(6):539–557View ArticleGoogle Scholar
- Intajag S, Tipsuwanporn V, Chatthai RC (2009) Retinal image enhancement in multi-mode histogram. In: World congress on computer science and information engineering, vol 4, pp 745–749Google Scholar
- Kumar A, Shaik F (2015) Image processing in diabetic related causes. Springer, Berlin. ISBN: 978-981-287-623-2Google Scholar
- Lu C, Mahmood M, Jha N, Mandal M (2012) A robust automatic nuclei segmentation technique for quantitative histopathological image analysis. Anal Quant Cytopathol Histpathol 34(6):296-308Google Scholar
- Peres FA, Oliveira FR, Neves LA, Godoy MF (2010) Automatic segmentation of digital images applied in cardiac medical images. In: IEEE-PACHE conference, workshops, and exhibits cooperation, Lima, Peru, March 15–19Google Scholar
- Ravindraiah R, Shaik F (2010) Detection of exudates in diabetic retinopathy images. In: National conference on future challenges and building intelligent techniques in electrical and electronics engineering (NCEEE ‘10), pp 363–368, Chennai, IndiaGoogle Scholar
- Shaik F, Giriprasad MN, Swathi C, Sekhar AS (2010) Detection Of cardiac complications in diabetic patients using Clahe method. In: Proceedings of the international conference on aerospace electronics, communications and instrumentation (ASECI 2010), pp 344–347, January 6–7, IndiaGoogle Scholar
- Sharifi A, Vosolipour A, Aliyari SM, Teshnehlab M (2008) Hierarchical Takagi–Sugeno type fuzzy system for diabetes mellitus forecasting. In: Proceedings of the 7th international conference on machine learning and cybernetics, Kunming, vol 3, pp 1265–1270, 12–15 JulyGoogle Scholar
- Sopharak A, Uyyanonvara B, Barman S, Williamson TH (2008) Automatic detection of diabetic retinopathy exudates from non-dilated retinal images using mathematical morphology methods. Comput Med Imaging Graph 32:720–727View ArticleGoogle Scholar
- Tang L, Niemeijer M, Abramoff MD (2011) Splat feature classification: detection of the presence of large retinal hemorrhages. In: IEEE international symposium on biomedical imaging: from nano to macro, pp 681–684Google Scholar
- Zana F, Klein J-C (1999) A multimodal registration algorithm of eye fundus images using vessels detection and hough transform. IEEE Trans Med Imaging 18(5):419–428View ArticleGoogle Scholar