Diagnostic support apparatus and diagnostic support method

摘要

The diagnostic support apparatus includes a feature quantity calculation unit vectorizing a test image into shift-invariant feature quantities, a base representation unit transforming the shift-invariant feature quantities of the test image into a linear combination of base vectors with first coefficients, the base vectors being calculated from a plurality of vectors representing shift-invariant feature quantities of a plurality of normal structural images including no lesion site, a lesion determination unit determining that the test image includes a lesion site when a difference between the first coefficients and second coefficients is greater than a determination threshold value, the second coefficients being used to transform shift-invariant feature quantities calculated from one of the normal structural images into a linear combination of the base vectors, and an output unit outputting a result of the determination by the lesion determination unit.

主权项

1. A diagnostic support apparatus comprising:
a non-transitory memory that stores a program; and a hardware processor that executes the program and causes the diagnostic support apparatus to operate as: a feature quantity calculation unit configured to vectorize a test image into shift-invariant feature quantities, the test image being an image in which presence of a lesion site is unknown; a base representation unit configured to transform the shift-invariant feature quantities of the test image into a linear combination of base vectors with first coefficients, the base vectors being represented by eigen vectors calculated through principal component analysis of a plurality of vectors representing shift-invariant feature quantities of a plurality of normal structural images that include no lesion site; a storage unit in which second coefficients are stored, the second coefficients being used to transform shift-invariant feature quantities calculated from one of the plurality of normal structural images into a linear combination of the base vectors; a lesion determination unit configured to determine that the test image includes a lesion site when a difference between the first coefficients obtained from a result of the transforming by the base representation unit and the second coefficients stored in the storage unit is greater than a determination threshold value; and an output unit configured to output a result of the determination by the lesion determination unit.