Segmentation of Food Items Using Watershed Algorithm and Predicting the Country of Food Items
A.D. Anantha Padmanabha Reddy1, P. Sriramya2
1A.D. Anantha Padmanabha Reddy, UG Scholar, Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences, Thandalam, Chennai (Tamil Nadu), India.
2P. Sriramya, Associate Professor, Faculty of Engineering, Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences, Thandalam, Chennai (Tamil Nadu), India.
Manuscript received on 27 April 2019 | Revised Manuscript received on 09 May 2019 | Manuscript Published on 17 May 2019 | PP: 497-501 | Volume-7 Issue-6S4 April 2019 | Retrieval Number: F11030476S419/2019©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Performing segmentation on images is a challenging task and interesting task in the field of image processing. Nowadays segmenting the intake of food for every meal and classifying them has become a challenge for user. It is important to assess the food that is taken by people, patients so that they can take of their diet when they fall under any internal diseases. In addition, there exists a problem of eating various kinds of food usually different country food, which actually decreases the resistance of the body etc. Hence, there exists the need to segment the food items and classifying the food items based on country. In addition, the advancement of this can be deployed to evaluate the nutrition content by importing equations in future. However, many existing systems discuss on how to develop an efficient dietary management system and nutrition estimation using various models and algorithms. The food retrieval and classification plays a vital role in every food based dietary management system. In the proposed work, we implement Watershed Algorithm to categorise from a bunch of sample food digital images by performing segmentation analysis and finally displays the unique number of segments whenever there is an overlapping in the food image. In addition, the classification model displays the country with their accuracy through image classification methods. The experimental results shows the accuracy of the classifier model in python and keras tools when different number of epochs are used for the model.
Keywords: Classification, Segmentation, Watershed Algorithm, Identification.
Scope of the Article: Algorithm Engineering