公开(公告)号：EP3932292A1

公开(公告)日：2022-01-05

申请号：EP19916685.1

申请日：2019-12-24

Applicant: TOPCON CORPORATION

Inventor： HIROSE Ryoichi ,  YAMAGUCHI Tatsuo ,  MIYAGAWA Suguru

IPC: A61B3/10

Abstract: This ophthalmic information processing device analyzes an image of a test eye formed by arranging multiple A scan images obtained by OCT scanning inside of the test eye with a measurement beam deflected about a scan center position. This ophthalmic information processing device includes a correction unit, a region specifying unit and a direction specifying unit. The correcting unit converts pixel positions in the aforementioned image to converted positions that are along the direction of advancement of the measurement beam which passes through the scan center position. The region specifying unit specifies a prescribed layer region by analyzing the image in which the pixel positions have been converted by the correcting unit. The direction specifying unit specifies the normal direction of the layer region specified by the region specifying unit.

公开(公告)号：EP3531068A1

公开(公告)日：2019-08-28

申请号：EP18760902.9

申请日：2018-02-28

Applicant: Riken ,  Topcon Corporation

Inventor： YOSHIZAWA Shin ,  YOKOTA Hideo ,  MIYAGAWA Suguru

IPC: G01B21/00 ,  G01C7/04

Abstract: [Problem] To easily and automatically extract analysis-target point cloud data as to a structure as an analysis target from point cloud data as to a measurement route acquired over the entire perimeter of a measurement device.
[Solution] Target point cloud data as point cloud data as to a specific road is extracted from entire perimeter point cloud data acquired by moving a road surface measurement device 300 along a measurement route and scanning the perimeter of the device. An extraction device 100 stores, in a data storage unit 110, the entire perimeter point cloud data and locus point string data representing a locus of movement of the road surface measurement device 300 on the measurement route as a plurality of locus points, acquires, by a locus point string setting unit 120, a locus point string equally spaced from the locus point string data, sets, by an extraction area setting unit 130, as an extraction area, a columnar area Ci and a parallelepiped area Hi each as a geometrical area arranged at a predetermined position below a locus point Xi, and extracts, by an approximate nearest neighbor processing unit 140 and an extraction processing unit 150, as target point cloud data, point data in the entire perimeter point cloud data belonging to this extraction area.