© Consumes your monthly credits. Limited amount of monthly calls, exceeding calls will be charged as extras.

(D) Limited amount of daily calls per user.

(S) Limited amount of calls per second.

This endpoint must be called attached to an already existent food segmentation prediction (Pre-consumption image). It applies the same food segmentation procedure as in /image/segmentation/complete but on an image that contains leftovers. For all the waste food item positions that were estimated to be matched to an item in the preconsumption image, the parameter 'serving_size' will be returned with the estimated food quantity.

Optionally, the endpoint allows providing a depth image in the same format as /image/segmentation/complete/quantity to increase the leftovers estimation precision. The depth image must be extracted using the LogMeal Depth SDK, that you will have to integrate into your App accordingly. Two SDKs are available, the iOS Depth SDK and the Android Depth SDK. If a depth image is provided an additional attribute named 'serving_size' is provided for each and every one of the elements inside the 'segmentation_results' list. Note that the segmentation_results might not be directly applicable to the uploaded image size. The segmentation pixel values are scaled depending on the 'processed_image_size' values provided for each of the images (the images might be sometimes resized depending on the needs of the food recognition algorithms used).

IMPORTANT: Note that this endpoint has two parameters in the request body ('image' and 'depth') which are in the 'multipart/form-data', and the remaining parameters are in 'application/json'. You can check it from the dropdown from the right corner.