| With the development of AI,AI has come into people's daily life.Diet is a part of human daily life,and thus the studies of AI in diet has become very popular.The field of food recognition is getting much attention.Food recognition is the basis of the researches of food,however,most food recognition takes aimed at Western food and Japanese food,while this paper proposes a food recognition method based on Chinese food.Since the deep learning was proposed,deep learning has made great achievements in many fields such as image recognition,object detection,natural language processing and so on.Deep learning is applied to this paper,where deep convolutional neural network model is utilized to detect and recognize Chinese food.The images are shared according to the detection results,the resulting main areas of Chinese food are retained in this way.A number of deep convolutional neural networks are integrated by the selective ensemble method,and then,the ensemble models are utilized to recognize Chinese food images which have been detected to achieve faster,more accurate and more stable Chinese food recognition.This paper uses the Chinese food detection method and the ensemble convolutional neural network models to recognize Chinese food,and applies many excellent deep convolutional neural network models to deal with the tasks of Chinese food recognition.During the task of Chinese food detection,the establishment of Chinese food detection database is a major work of this paper.The use of Chinese food detection and the implementation of using selective ensemble method to build integration models for the task of Chinese food recognition are another work in this paper.The experiment is conducted on the VIREO Food-172 database by combining the model of Chinese food detection and the ensemble model constructed by the selective ensemble method,which achieves a satisfactory recognition accuracy. |
