Information processing apparatus, learning method, recognition method, and non-transitory computer readable medium

A learning method for a learning model to which image data and region image data on a first region included in the image data are inputted so as to perform predetermined recognition on the input image data comprising:

• a first acquisition step of acquiring learning image data, ground truth region image data which indicates a ground truth region of the first region included in the learning image data, and ground truth data on the recognition,
• a second acquisition step of acquiring estimated region image data which indicates an estimated region of the first region in the learning image data;
• a learning step of training the learning model using a first training dataset including the learning image data to be input to the learning model, the ground truth region image data to be input to the learning model, and the ground truth data on the recognition;
• and
• further train the learning model, which has been trained using the first training dataset, using a second training dataset including the learning image data to be input to the learning model, the estimated region image data to be input to the learning model and the ground truth data on the recognition.

A learning method for a learning model to which image data and region image data on a first region included in the image data are inputted so as to perform predetermined recognition on the input image data comprising:

• a first acquisition step of acquiring learning image data, ground truth region image data which indicates a ground truth region of the first region included in the learning image data, and ground truth data on the recognition,
• a second acquisition step of acquiring estimated region image data which indicates an estimated region of the first region in the learning image data;
• and
• a learning step of training the learning model using a first training dataset constituted of the learning image data, the ground truth region image data, and the ground truth data, and a second training dataset constituted of the learning image data, the estimated region image data and the ground truth data,
• wherein, in the learning step, a third training dataset is constructed from training data included in the first training dataset and training data included in the second training dataset, and the learning model is trained using the third training dataset, and
• wherein the third training dataset is, in a case of satisfying a predetermined condition, constructed by changing the ratio of a number of training data included in the first training dataset and a number of training data included in the second training dataset.