This model is trained to recognize 2023 food dishes from images. It is based on MobileNet V1.
This model takes images as input.
Inputs are expected to be 3-channel RGB color images of size 224 x 224, scaled to [0, 1].
This model outputs to image_classifier.
image_classifier: A probability vector of dimension 2024, corresponding to a background class and one of 2023 food dishes in the labelmap.
For details of the model architecture, see MobileNet V1.
This model is trained to recognize 2023 food dishes. The training set includes entrees, side dishes, desserts, snacks, etc.
Do not use this model to determine whether an object is edible or not. This model is not suitable for predicting the ingredients of a food dish. Do not use this model to predict allergen or nutrition information.
This model assumes that its input image contains a well-cropped food dish. If a non-food image is input to the model or if the dish is not well-cropped, the output of the model may be meaningless. This model was trained on a dataset skewed toward North American foods.