Biological information processor
According to one embodiment, a biological information processor includes: an area detection unit configured to detect an area in which a person is displayed from image information; a feature extraction unit configured to extract feature information based on a characteristic portion of a person from the area detected by the area detection unit from the image information; and a determination unit configured to determine an extraction precision indicating whether or not the characteristic portion of the person can be extracted, with respect to the feature information extracted by the feature extraction unit based on a position of the characteristic portion.
1. A biological information processor comprising:
at least one processor configured
to detect an area in which a person is displayed from image information,
to extract, from the detected area, feature information which is based on a characteristic portion of a person,
to perturb a position of a characteristic point detected in test image information to set a perturbation feature point group indicating positions different from the position where the characteristic point is detected as feature points,
to extract second feature information from the perturbation feature point group,
to determine, based on the image information, the detected area and a position of the characteristic portion of the extracted feature information, an extraction precision of the feature information indicating whether the characteristic portion of the person is correctly extracted,
to calculate a similarity between the feature information and the second feature information indicating a position different from the position where the characteristic portion is detected as a feature,
to determine, if the similarity is equal to or larger than a predetermined threshold, that the extraction precision of the feature information indicates that the characteristic portion is not correctly extracted, and
to correct, if the extraction precision indicates that the characteristic portion is not correctly extracted, at least the detected area among the image information, the detected area, and the feature information without re-detecting the area and re-extracting the feature information so that the extraction precision is increased.
2. The biological information processor according to claim 1 , further comprising a storage configured to store prepared image information for comparison in which a person is displayed,
wherein the at least one processor is configured to generate normalized image information obtained by extracting the area in which the person is displayed from the image information with the position of the characteristic portion of the person as a reference, and compare the generated normalized image information with the image information for comparison stored in the storage to determine the extraction precision.
3. The biological information processor according to claim 1 , wherein the at least one processor is configured to calculate at least one of a centroid of the image information based on a plurality of the positions of a plurality of the characteristic portions of the person and a distance between the positions, and determine the extraction precision based on whether or not the at least one of the centroid and the distance calculated is higher than a predetermined threshold.
4. The biological information processor according to claim 1 , wherein the at least one processor is configured to extract feature information based on a position representing one of a face, a mouth, an iris, retina, an ear, a vein shape of a part of a human body, a fingerprint, and a palm print of the person included in the detected area.
5. The biological information processor according to claim 1 , wherein the image information is one of image information captured by an image capturing unit and the image information stored in a storage used for authentication of a person.
6. The biological information processor according to claim 1 , wherein the at least one processor is configured to display a screen for correction of the at least one of the image information, the detected area, and the feature information.
7. The biological information processor according to claim 1 , wherein
the at least one processor further calculates a first evaluation value based on the image information, a second evaluation value based on the detected area and the position of the characteristic portion of the extracted feature information, and a third evaluation value based on the position of the characteristic portion of the extracted feature information,
the at least one processor is configured to determine the extraction precision of the feature information based on the first evaluation value, the second evaluation value, and the third evaluation value, and
the at least one processor is configured to correct, if the extraction precision indicates that the characteristic portion is not correctly extracted, the at least the detected area among the image information, the detected area, and the feature information without re-detecting the area and re-extracting the feature information so that at least one of the first evaluation value, the second evaluation value, and the third evaluation value is increased.