Abstract: A processor is configured to extract a target organ from a medical image, set a plurality of small regions in the target organ, derive a first evaluation value representing a physical quantity of each of the plurality of small regions, derive at least one second evaluation value representing a relationship between the first evaluation values in the plurality of small regions, and derive a third evaluation value indicating presence or absence of an abnormality in the entire target organ based on the second evaluation value.

Abstract: A processor is configured to set a first region including an entire target organ for a medical image, set a plurality of small regions including the target organ in the first region, derive a first evaluation value indicating presence or absence of an abnormality in the first region, derive a second evaluation value indicating the presence or absence of the abnormality in each of the plurality of small regions, and derive a third evaluation value indicating the presence or absence of the abnormality in the medical image from the first evaluation value and the second evaluation value.

Abstract: A processor is configured to extract a tubular tissue region from a medical image including a tubular tissue, extract a lesion region in a periphery of the tubular tissue region from the medical image, and correct an extraction result of the tubular tissue region, as necessary, based on positional information of the lesion region.

Abstract: A processor derives a feature amount related to an input image based on first to nth (n?2) derivation models, derives, via the first derivation model, a first feature amount based on the input image, and derives, via a kth (1<k?n) derivation model, a kth feature amount based on the input image and a k?1th feature amount derived by a k?1th derivation model. An nth feature amount derived by the nth derivation model is the feature amount related to the input image, and the k?1th feature amount includes a feature amount related to another region that is independent of a region related to the kth feature amount, which suggests the kth feature amount derived by the kth derivation model, and that serves as a basis for the derivation of the kth feature amount.

Abstract: Provided are a medical image processing device, an operation method of the medical image processing device, and an operation program of the medical image processing device, which can contribute to find the lesion that is hardly depicted, by using a method based on the thought of the doctor. A CPU includes an image acquisition unit, a first derivation unit, and a second derivation unit. The image acquisition unit acquires a medical image. The first derivation unit analyzes the medical image to derive indirect finding information related to an indirect finding that represents at least one feature of a shape or a property of peripheral tissue of the lesion associated with an occurrence of the lesion. The second derivation unit derives lesion presence probability information indicating a presence probability of the lesion based on the indirect finding information.