International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395 -0056
Volume: 04 Issue: 04 | Apr -2017
p-ISSN: 2395-0072
www.irjet.net
Tongue Image Analysis with Color and Gist Features for Diabetes Diagnosis. Balkrishna B. Gabhale1, Meenal S. Shinde2, Ashwini M. Kamble3, Mrs. V.C.Kulloli4 1Student,
Information Technology, Pimpri Chinchwad College of Engineering, Maharashtra, India. Student, Information Technology, Pimpri Chinchwad College of Engineering, Maharashtra, India. 3 Student, Information Technology, Pimpri Chinchwad College of Engineering, Maharashtra, India. 4 Assistant Professor, Information Technology, Pimpri Chinchwad College of Engineering, Maharashtra, India. --------------------------------------------------------------------------***----------------------------------------------------------------the actual functionality of organs. For example, dark red Abstract - The tongue is an important part of human body to 2
taste, speak and swallow the food. It reflect the inner working of body. Any kind of unusual behavior of body is reflected through tongue like problem in stomach, pancreas, liver and intestines etc. So in Ayurveda and modern science, doctors initially inspect the tongue based on its color, texture and geometry to diagnose the disease. Changes in various features of tongue easily predict the misbehavior in human body. So in this paper, we are proposing automatic computer based technique to analyze the changes in tongue, which will be latter useful in diabetes diagnosis in patients. So initially we take the images of tongue and apply the image processing based feature extraction technique to extract the two types of features named as, Color features and Gist features. Latter on these quantifies features will be classify by using ANN based SOM Kohonen Classifier. To examine the performance of system, we are using tongue image dataset. Experimental results will prove that the color and gist features provides more accuracy in classification results than color and texture features .
color indicate the inflammation or ulceration, a white color indicates cold attacker a weakness in the blood. This attributes leads to conditions of anemia. On the other side, a yellow tongue presents the diseases related to liver and gallbladder, and blue or purple related to difficulties in blood circulation or a weakness in digestive system. In healthy tongue, color is a nice, robust, sanguine(reddish) pink [2].
Key Words: SOM kohonen, tongue image analysis, color features, gist features, diabetes diagnosis
Changes in various features of tongue easily predict the misbehavior in human body. So in this proposed paper, we are developing automatic computer based technique to analyze the changes in tongue, which will be latter useful in diabetes diagnosis in patients. Here, computer based tongue analysis method is proposed. This method extracts the quantitative color and gist features (low dimensional representation of the scene, which does not require any form of segmentation) of images of tongue. Then system apply the SOM Kohonen classifier on these extracted quantitative features to classify the diagnosis of diabetes. SOM Kohonen [1] is one the popular type of Artificial Neural Network classifier and two types of features namely color and gist measures are extracted from images by using digital image processing technique. Finally classifier maps the relationship between features and disease.
1.INTRODUCTION The tongue can be used for the diagnosis of disease by using its various features such as color, texture, geometry etc. Figure 1 represent the various sides of tongue reflecting the behaviors of various part of human body. Tip of the tongue is related to lungs, heart, chest and neck. The central part is related to liver, spleen, stomach and pancreas. The rear one reflects the lower abdominal organs – small intestine and the colon(lagre Intestine). So to detect the diseases, system has to inspect the particular area of tongue. That is the infected area of tongue reflects the misbehavior in particular respected part of body. This paper concentrate on the diagnosis of Diabetes. For this purpose we are extracting the features form all tongue image specifically from middle region. Because middle region is related to the pancreas. We extract the two features from region like color and gist feature. Most attributes are considers for diagnosis such as color, moisture, size, shape, and coating. These attributes reveals Š 2017, IRJET
|
Impact Factor value: 5.181
|
Fig- 1. Structure of Tongue
2.LITERATURE SURVEY Paper [3] studies and analyzes the tongue shape to establish a kind of tongue diagnosis method based on tongue images. CCD devices were employed to acquire the front and lateral images of tongue, and then measure the tongue's length (L), width (K) and height (H). Furthermore, in 500 ISO 9001:2008 Certified Journal
|
Page 523