Food is the fuel of human body & one of the basic necessities of human beings. Due to modern life style dietary habits of human being have changed which include consumption of ready mode, packaged & fast food with the reduction of physical labour or exercise carried out by human beings. This kind of unbalanced diet is a high risks factor for diseases & ailments such as obesity, cardiac problems & a host of other diseases. Our work is aimed at determination or classification of food using image processing in conjunction with other intelligent algorithms, with the ultimate aim of determination/estimation of calorie intake our work acts as basis of modern computer assisted, remote dietary management systems. Our system comprises of segmentation of food in the image, then extracting image parameters such as area, major axis, minor axis convex area from the segmented food area, & then using an already trained artificial neural network to classify the food on basis of these parameters. Multiple methods have been combined using weighted averaging to achieve food segmentation, such as surface feature/ bag of features detection; background removed using HCV processing etc. High detection accuracy is obtained by combination of multiple image processing techniques with leven barg marquard function flitting neural network.
Keywords : Food Recognition, SURF Based Bag of Features, HSV Background Elimination, Image Based Calorie Estimation, Leven Barg Marquard Neural Network.