Apparatus, method, and computer program product for medical diagnostic imaging assistance

A method for medical diagnostic imaging assistance for automatically detecting a lesion on a target case image using an apparatus for medical diagnostic imaging assistance, the apparatus including a sampling unit, a machine-learning unit, a statistical processing unit, a first detection unit, a first extraction unit, and a determination unit, the method comprising:

• sampling, based on a first sampling information, from a memory storing first feature information representing the feature of a lesion mask or a non-lesion mask as first learning data per mask to acquire a plurality of samples of more than one of the first learning data, and generating a plurality of different subsets of learning data;
• generating a first discrimination condition corresponding to each of the plurality of different subsets by carrying out a machine-learning processing step on the plurality of different subsets in the machine-learning unit;
• generating a second discrimination condition by carrying out a statistical processing step under the first distribution condition in the statistical processing unit;
• wherein category information indicating a category used for medical or statistical classification is added to the first learning data stored in the memory,
• wherein the first learning data are sampled and the plurality of different subsets of learning data are generated based on the category information, and
• wherein a number of samples of the first learning data is determined based on the category information.

A non-transitory computer readable medium storing a program for allowing a computer to execute an automatic detection process for detecting a lesion on a target case image, the process comprising:

• acquiring a plurality of samples by sampling from a memory storing first feature information representing a lesion mask or a non-lesion mask;
• generating a first discrimination condition corresponding to each of the samples by carrying out a machine-learning processing step on the plurality of samples;
• generating a second discrimination condition by carrying out a statistical processing step under the first discrimination condition;
• wherein category information indicating a category used for medical or statistical classification is added to first learning data stored in the memory,
• wherein the first learning data are sampled and a plurality of different subsets of learning data are generated based on the category information, and
• wherein a number of samples of the first learning data is determined based on the category information.