Shape recognition device, shape recognition program, and shape recognition method
First Claim
1. A shape recognition device comprising:
- a depth level sensor that measures a distance to an object and detects an external shape of a hand of a human body according to measured distance; and
a control unit that comprises a memory and a processor and performs image recognition functions that recognize fingers of the hand, the image recognition functions comprising;
a polygon creation function that creates a polygon having a vertex at each of a plurality of external-shape points based on the external shape;
a tip point extraction function that extracts a tip point of each finger from common vertexes between the polygon and a convex hull of the polygon;
a vector calculation function that calculates a vector of a reference line segment that connects a pair of right and left vertexes of the polygon adjacent to the tip point and a vector of each side that constitutes the polygon;
a base point extraction function that extracts a base point of each finger based on a position of a side which is most parallel to the reference line segment; and
an extending-direction detecting function that extracts a skeleton of each finger by connecting the tip point and the base point of each finger, and calculates a direction of the skeleton.
3 Assignments
0 Petitions
Accused Products
Abstract
A shape recognition device, a shape recognition program, and a shape recognition method that can obtain more precise information for recognizing an external shape of an object are provided. A shape recognition device of the present invention includes an external-shape detecting unit that detects an external shape of an object, a first extracting unit that extracts a first feature point on the external shape based on the external shape detected by the external-shape detecting unit, and a second extracting unit that extracts a second feature point that exists on the external shape or in the external shape and differs from the first feature point. As a result, more precise information for recognizing the external shape of the object can be obtained.
15 Citations
20 Claims
-
1. A shape recognition device comprising:
-
a depth level sensor that measures a distance to an object and detects an external shape of a hand of a human body according to measured distance; and a control unit that comprises a memory and a processor and performs image recognition functions that recognize fingers of the hand, the image recognition functions comprising; a polygon creation function that creates a polygon having a vertex at each of a plurality of external-shape points based on the external shape; a tip point extraction function that extracts a tip point of each finger from common vertexes between the polygon and a convex hull of the polygon; a vector calculation function that calculates a vector of a reference line segment that connects a pair of right and left vertexes of the polygon adjacent to the tip point and a vector of each side that constitutes the polygon; a base point extraction function that extracts a base point of each finger based on a position of a side which is most parallel to the reference line segment; and an extending-direction detecting function that extracts a skeleton of each finger by connecting the tip point and the base point of each finger, and calculates a direction of the skeleton. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8)
-
-
9. A shape recognition system comprising:
-
a depth level sensor measures a distance to an object and detects an external shape of a hand having fingers of a human body according to measured distance; a memory to store computer executable instructions; and a processor to execute the computer executable instructions to perform processes, the processes comprising; a polygon creation process of creating a polygon having a vertex at each of a plurality of external-shape points based on the external shape; a tip point extraction process of extracting a tip point of each finger of the hand having fingers from common vertexes between the polygon and a convex hull of the polygon; a vector calculation process of calculating a vector of a reference line segment that connects a pair of right and left vertexes of the polygon adjacent to the tip point and a vector of each side that constitutes the polygon; a base point extraction process of extracting a base point of each finger based on a position of a side which is most parallel to the reference line segment; and an extending-direction detecting process of extracting a skeleton of each finger by connecting the tip point and the base point of each finger, and calculating a direction of the skeleton. - View Dependent Claims (10, 11, 12, 13, 14)
-
-
15. A shape recognition method comprising:
-
a shape detecting step of detecting an external shape of a hand having fingers of a human body using data of distance to an object measured by a depth level sensor; a polygon creation step of creating a polygon having a vertex at each of a plurality of external-shape points based on the external shape; a tip point extraction step of extracting a tip point of each finger of the hand having fingers from common vertexes between the polygon and a convex hull of the polygon; a vector calculation step of calculating a vector of a reference line segment that connects a pair of right and left vertexes of the polygon adjacent to the tip point and a vector of each side that constitutes the polygon; a base point extraction step of extracting a base point of each finger based on a position of a side which is most parallel to the reference line segment; and an extending-direction detecting step of extracting a skeleton of each finger by connecting the tip point and the base point of each finger, and calculating a direction of the skeleton. - View Dependent Claims (16, 17, 18, 19, 20)
-
Specification