Apparatus which estimates inside of object and method of estimating inside of object

Patent metadata

Field	Value
Publication number	US-9234747-B2
Application number	US-201313959842-A
Country	US
Kind code	B2
Filing date	Aug 6, 2013
Priority date	Aug 13, 2012
Publication date	Jan 12, 2016
Grant date	Jan 12, 2016

Abstract

An apparatus which estimates an inside of an object, includes: a first obtaining unit which obtains light path models and a first light intensity distribution, based on an assumption that a first object includes no second object, distribution; a second obtaining unit which obtains a second light intensity distribution which is an intensity distribution of observed light; a voting value calculating unit which calculates a voting value for each of positions on a predetermined plane, based on a result of comparison in light intensity between the first and second light intensity distributions; a voting unit which votes the voting value to voting regions on a light path model along which light reaches each position; and an estimating unit which estimates whether the second object exists in each voting region, based on a result of the voting.

First claim

The invention claimed is:

1. An apparatus which estimates an inside of a first object, the apparatus comprising:
a non-transitory memory storing a program;
a hardware processor that executes the program and causes the apparatus to obtain a plurality of light path models and a first light intensity distribution, based on an assumption that the first object includes no second object which is different from the first object, the light path models representing a plurality of paths taken by light after entering the first object and before exiting the first object, the first light intensity distribution indicating an intensity of virtual light on a predetermined plane, the virtual light exiting the first object after propagating through the first object along the light path models, the first light intensity distribution being indicated by light transmittance derived from the light path models obtained as a result of ray tracing simulation; and
a camera configured to obtain a second light intensity distribution which is an intensity distribution of light exiting the first object after entering the first object, a light source emitting visible light or near-infrared light to the object and the camera obtaining an observation image indicating the second light intensity distribution;
wherein the hardware processor executes the program and causes the apparatus to
calculate a voting value for each of positions on the predetermined plane, based on a result of comparison in light intensity between the first light intensity distribution and the second light intensity distribution, the voting value indicating a probability of existence of the second object in the first object;
virtually divide the inside of the first object into a plurality of voting regions, and add the voting value calculated for the each of the positions on the predetermined plane to values of voting regions on a light path model along which light reaches the position, the adding being performed for the each of the positions; and
estimate whether or not the second object exists in each of the voting regions, by comparing a result of the adding of the voting value for each of the voting regions with a threshold.

2. The apparatus according to claim 1 , wherein the hardware processor executes the program and causes the apparatus to:
correct the voting result, based on a variation in a total number of times the voting was performed among the voting regions, and
estimate whether or not the second object exists in each of the voting regions, based on the corrected voting result.

3. The apparatus according to claim 1 ,
wherein the voting regions are a plurality of regions on a cross-section of the first object,
the estimating generates a plurality of two-dimensional tomographic images by estimating, for each of a plurality of the cross-sections, whether or not the second object exists in each of the voting regions, and
the hardware processor executes the program and causes the apparatus to
generate three-dimensional volume data of the inside of the first object, by combining the two-dimensional tomographic images.

4. The apparatus according to claim 1 ,
wherein the camera includes:
a light source unit configured to emit visible light or near-infrared light to the first object; and
a capturing unit configured to obtain an observation image indicating the second light intensity distribution, by capturing an image of light exiting the first object after entering the first object.

5. The apparatus according to claim 1 ,
wherein the hardware processor executes the program and causes the apparatus to derive the light path models by Monte Carlo ray tracing simulation.

6. The apparatus according to claim 5 ,
wherein the hardware processor executes the program and causes the apparatus to derive a parameter to be used in the Monte Carlo ray tracing simulation, by reducing an error between a simulation result and an observation result through a coarse-to-fine search.

7. The apparatus according to claim 1 ,
wherein the hardware processor executes the program and causes the apparatus to obtain the light path models and the first light intensity distribution for each of a plurality of light beams which enters the first object from different directions,
the camera obtains the second light intensity distribution for each of the light beams,
the hardware processor executes the program and causes the apparatus to calculate the voting value for each of the light beams, and
vote the voting value for each of the light beams.

8. The apparatus according to claim 1 , wherein the hardware processor executes the program and causes the apparatus to
correct the voting result, based on the position of each of the voting regions, and
estimate, based on the corrected voting result, whether or not the second object exists in each of the voting regions.

9. The apparatus according to claim 8 ,
wherein the voting regions include a first voting region and a second voting region that is closer to a center of the first object than the first voting region is, and
the hardware processor executes the program and causes the apparatus to correct the voting result to allow the probability of existence of the second object in the first voting region to be relatively lower than the probability of existence of the second object in the second voting region.

10. A method for an apparatus of estimating an inside of a first object, the apparatus including a camera, a non-transitory memory storing a program and a hardware processor that executes the program and causes the apparatus to perform the method comprising:
obtaining, using the hardware processor, a plurality of light path models and a first light intensity distribution, based on an assumption that the first object includes no second object which is different from the first object, the light path models representing a plurality of paths taken by light after entering the first object and before exiting the first object, the first light intensity distribution indicating an intensity of virtual light on a predetermined plane, the virtual light exiting the first object after propagating through the first object along the light path models, the first light intensity distribution being indicated by light transmittance derived from the light path models obtained as a result of ray tracing simulation;
obtaining, using the camera, a second light intensity distribution which is an intensity distribution of light exiting the first object after entering the first object, a light source emitting visible light or near-infrared light to the object and the camera obtaining an observation image indicating the second light intensity distribution;
calculating, using the hardware processor, a voting value for each of positions on the predetermined plane, based on a result of comparison in light intensity between the first light intensity distribution and the second light intensity distribution, the voting value indicating a probability of existence of the second object in the first object;
virtually dividing, using the hardware processor, the inside of the first object into a plurality of voting regions, and adding the voting value calculated for the each of the positions on the predetermined plane to values of voting regions on a light path model along which light reaches the position, the adding being performed for the each of the positions; and
estimating, using the hardware processor, whether or not the second object exists in each of the voting regions, by comparing a result of the adding of the voting value for each of the voting regions with a threshold.

Assignees

• Panasonic Corp

Inventors

• Ishii Yasunori
• Arai Toshiya
• Mukaigawa Yasuhiro
• Yagi Yasushi

Classifications

• G01B11/24 Primary

  for measuring contours or curvatures · CPC title