Malay festive seasons food recognition for calorie detection using SVM and ECOC approaches / Nurul Hafiza Binti Basiruddin, Zalikha Zulkifli and Samsiah Ahmad

The idea of adding an auto-recognition feature for Malay Festive Seasons Food based on images is a very challenging task in computer vision as it is something new and undiscovered until recently. However, this recognition is important for Malaysian users to manage calorie intake, especially during Hari Raya, one of Malaysia's biggest festive seasons and most celebrated festivals. As color plays an important role in differentiating the type of food, this research aims to implement Color Feature Extraction Method after performing segmentation techniques during the pre-processing phase, where each color from the images is extracted individually. Then the result from the Color Feature Extraction Method is used to identify the type of food by using Error-Correcting Output Codes (ECOC) classification, which is part of the Support Vector Machine (SVM) algorithm. The reliability and effectiveness of the classifier are evaluated through system testing, where the total overall percentage of correct recognition performed by the system is 82.5%, according to the correct and wrong recognition obtained. The ability to recognize the food correctly after classifying the image is crucial in this research to accurately perform the calorie estimation, whereby the calorie value will be auto generated after food recognition is performed.