2019 · 16 citations · 10 references
ComputingComputer ScienceComputer VisionFood AnalysisMachine LearningFeature DetectionImage ClassificationObject DetectionObject RecognitionTexture AnalysisColor SpacesGray-level Co-occurrence MatricesHsv Color SpaceLocal Binary PatternsRandom Forest ClassificationTexture Feature DetectionTexture Feature ExtractionTexture Features
Indonesia has various cultures and consequently, it also has huge selection of traditional foods. Since there is new style of people of sharing food images in several social media before eating nowadays, people also have benefit to explore more information, such as recommendation place to eat particular food. When people travel to one place to another place they also have an intention to enjoy the local food, but some of them may not acknowledge the name of traditional food, especially for those who is not a residence in that place. By spreading images to others, it can help other people finding recommended food when visiting the same place. One way to tackle this problem is by identifying food images automatically. In this paper, we propose a new method for recognizing traditional foods in Indonesia. There are 382 images taken using smartphone camera, consisting of 33 classes as training set and 33 images as testing set. All features are extracted using color moments in RGB, LAB, and HSV color spaces and texture feature of Local Binary Pattern (LBP) and Gray Level Co-occurrence Matrix (GLCM). Random Forest Classifier with different number of estimators are used to evaluate the proposed method. The result of combined features consisting of color and texture features reaches 93.5% of accuracy using 500 estimators. It proves that this method is sufficient to recognize traditional food image.
10
Recipe recognition with large multimodal food dataset
Xin Wang, Devinder Kumar, Nicolas Thome et al. · 2015 · 179 citations · Full text
Doyen Sahoo, Hao Wang, Shu Ke et al. · 2019 · 99 citations · Full text
Trees Weighting Random Forest Method for Classifying High-Dimensional Noisy Data