Extremal Region Selection for MSER Detection in Food Recognition

The visual analysis of foods on social media by using food recognition algorithm provides valuable insight from the health, cultural and marketing. Food offers a means to automatically recognise as well useful information such calories nutritional estimation image processing machine learning technique. interest points in can be detected effectively Maximally Stable Extremal Region (MSER). As MSER used global segmentation many images have complex background, there are numerous irrelevant detected. These considered noises that lead computation burden overall process. Therefore, this research proposes an Selection (ERS) improve detection reducing number extremal regions unsupervised based k-means algorithm. performance ERS is evaluated classification metrics rate (CR), error (ERT), precision (Prec.) recall (rec.) produced ERS. UECFOOD-100 UNICT-FD1200 two datasets benchmark proposed results found optimum parameters thresholds, able reduce with sustained performance.