| 摘要 |
An alignment apparatus includes: a voting unit that generates multiple pairs of a first feature selected from multiple features representing a first target object represented on one image, and a second feature selected from multiple features representing a second target object represented on the other image, obtains, for each of the pairs, a first translation amount for aligning the first feature with the second feature, and votes, in a coordinate system with coordinates each representing a translation amount, a voting value for the coordinate corresponding to the first translation amount; and a translation amount calculating unit that calculates the total of the voting values for each of the coordinates in the coordinate system, and sets the translation amount corresponding to the coordinate with the largest total of the voting values, as a second translation amount for aligning the first target object with the second target object. |
