Techniques for improved image disparity estimation
원문보기
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
G06K-009/00
H04N-013/02
G06K-009/32
G06T-005/00
H04N-005/247
G06T-007/00
H04N-013/00
출원번호
US-0630469
(2015-02-24)
등록번호
US-9900580
(2018-02-20)
발명자
/ 주소
Ma, Tao
Wu, Yi
Nestares, Oscar
Seshadrinathan, Kalpana
Sun, Wei
출원인 / 주소
Intel Corporation
인용정보
피인용 횟수 :
0인용 특허 :
16
초록▼
Techniques for improved image disparity estimation are described. In one embodiment, for example, an apparatus may comprise a processor circuit and an imaging management module, and the imaging management module may be operable by the processor circuit to determine a measured horizontal disparity fa
Techniques for improved image disparity estimation are described. In one embodiment, for example, an apparatus may comprise a processor circuit and an imaging management module, and the imaging management module may be operable by the processor circuit to determine a measured horizontal disparity factor and a measured vertical disparity factor for a rectified image array, determine a composite horizontal disparity factor for the rectified image array based on the measured horizontal disparity factor and an implied horizontal disparity factor, and determine a composite vertical disparity factor for the rectified image array based on the measured vertical disparity factor and an implied vertical disparity factor. Other embodiments are described and claimed.
대표청구항▼
1. At least one non-transitory machine-readable medium comprising a plurality of instructions that, in response to being executed on a computing device, cause the computing device to: determine a first measured disparity factor along a first dimension for a rectified image array comprising a plurali
1. At least one non-transitory machine-readable medium comprising a plurality of instructions that, in response to being executed on a computing device, cause the computing device to: determine a first measured disparity factor along a first dimension for a rectified image array comprising a plurality of rectified images by a first process arranged to: iteratively select a first candidate value;measure a first error associated with the selected first candidate value; andselect as the first measured disparity factor a first candidate value with which a minimized first error is associated; anddetermine a second measured disparity factor along a second dimension for the rectified image array by a second process arranged to: iteratively select a second candidate value;measure a second error associated with the selected second candidate value; andselect as the second measured disparity factor a second candidate value with which a minimized second error is associated, the minimized first error comprising a minimum value of a first pixel matching error function, and the minimized second error comprising a minimum value of a second pixel matching error function. 2. The at least one non-transitory machine-readable medium of claim 1, the first dimension comprising a horizontal dimension extending along a row of the rectified image array, and the at least one non-transitory machine-readable medium comprising instructions that, in response to being executed on the computing device, cause the computing device to: determine a measured vertical disparity factor for the rectified image array by a third process arranged to: iteratively select a vertical candidate value;measure a vertical error associated with the selected vertical candidate value; andselect as the measured vertical disparity factor a vertical candidate value with which a minimized vertical error is associated;determine an implied horizontal disparity factor for the rectified image array based on the measured vertical disparity factor; anddetermine an implied vertical disparity factor for the rectified image array based on the first measured disparity factor. 3. The at least one non-transitory machine-readable medium of claim 2, comprising instructions that, in response to being executed on the computing device, cause the computing device to: determine a composite horizontal disparity factor for the rectified image array based on the first measured disparity factor and the implied horizontal disparity factor; anddetermine a composite vertical disparity factor for the rectified image array based on the measured vertical disparity factor and the implied vertical disparity factor. 4. The at least one non-transitory machine-readable medium of claim 3, the composite horizontal disparity factor and the composite vertical disparity factor comprising integer pixel disparities. 5. The at least one non-transitory machine-readable medium of claim 4, comprising instructions that, in response to being executed on the computing device, cause the computing device to: determine a horizontal sub-pixel disparity factor and a vertical sub-pixel disparity factor for the rectified image array by a fourth process arranged to: iteratively select a sub-pixel candidate value pair comprising a horizontal sub-pixel candidate value and a vertical sub-pixel candidate value;measure a joint error associated with the selected sub-pixel candidate value pair;determine as the horizontal sub-pixel disparity factor a horizontal sub-pixel candidate value comprised within a sub-pixel candidate value pair with which a minimized joint error is associated; anddetermine as the vertical sub-pixel disparity factor a vertical sub-pixel candidate value comprised within the sub-pixel candidate value pair with which the minimized joint error is associated. 6. The at least one non-transitory machine-readable medium of claim 1, the first pixel matching error function and the second pixel matching error function comprising sums of pairwise block matching error functions. 7. The at least one non-transitory machine-readable medium of claim 1, the first dimension comprising a horizontal dimension extending along a row of the rectified image array, and the at least one non-transitory machine-readable medium comprising instructions that, in response to being executed on the computing device, cause the computing device to: determine a measured vertical disparity factor for the rectified image array by a third process arranged to: iteratively select a vertical candidate value;measure a vertical error associated with the selected vertical candidate value; andselect as the measured vertical disparity factor a vertical candidate value with which a minimized vertical error is associated;determine the first measured disparity factor for a longest horizontal baseline of the rectified image array; anddetermine the measured vertical disparity factor for a longest vertical baseline of the rectified image array. 8. The at least one non-transitory machine-readable medium of claim 1, one or both of the first measured disparity factor or the measured vertical disparity factor comprising a number of pixels. 9. An apparatus, comprising: a processor circuit; andan imaging management module to: determine a first measured disparity factor along a first dimension for a rectified image array comprising a plurality of rectified images by a first process arranged to: iteratively select a first candidate value;measure a first error associated with the selected first candidate value; andselect as the first measured disparity factor a first candidate value with which a minimized first error is associated; anddetermine a second measured disparity factor along a second dimension for the rectified image array by a second process arranged to:iteratively select a second candidate value;measure a second error associated with the selected second candidate value; andselect as the second measured disparity factor a second candidate value with which a minimized second error is associated the minimized first error comprising a minimum value of a first pixel matching error function, and the minimized second error comprising a minimum value of a second pixel matching error function. 10. The apparatus of claim 9, the first dimension comprising a horizontal dimension extending along a row of the rectified image array, and the imaging management module to: determine a measured vertical disparity factor for the rectified image array by a third process arranged to: iteratively select a vertical candidate value;measure a vertical error associated with the selected vertical candidate value; andselect as the measured vertical disparity factor a vertical candidate value with which a minimized vertical error is associated;determine an implied horizontal disparity factor for the rectified image array based on the measured vertical disparity factor; anddetermine an implied vertical disparity factor for the rectified image array based on the first measured disparity factor. 11. The apparatus of claim 10, the imaging management module to: determine a composite horizontal disparity factor for the rectified image array based on the first measured disparity factor and the implied horizontal disparity factor; anddetermine a composite vertical disparity factor for the rectified image array based on the measured vertical disparity factor and the implied vertical disparity factor. 12. The apparatus of claim 11, the composite horizontal disparity factor and the composite vertical disparity factor comprising integer pixel disparities. 13. The apparatus of claim 12, the imaging management module to: determine a horizontal sub-pixel disparity factor and a vertical sub-pixel disparity factor for the rectified image array by a fourth process arranged to: iteratively select a sub-pixel candidate value pair comprising a horizontal sub-pixel candidate value and a vertical sub-pixel candidate value;measure a joint error associated with the selected sub-pixel candidate value pair;determine as the horizontal sub-pixel disparity factor a horizontal sub-pixel candidate value comprised within a sub-pixel candidate value pair with which a minimized joint error is associated; anddetermine as the vertical sub-pixel disparity factor a vertical sub-pixel candidate value comprised within the sub-pixel candidate value pair with which the minimized joint error is associated. 14. The apparatus of claim 9, the first pixel matching error function and the second pixel matching error function comprising sums of pairwise block matching error functions. 15. The apparatus of claim 9, the first dimension comprising a horizontal dimension extending along a row of the rectified image array, and the imaging management module to: determine a measured vertical disparity factor for the rectified image array by a third process arranged to: iteratively select a vertical candidate value;measure a vertical error associated with the selected vertical candidate value; andselect as the measured vertical disparity factor a vertical candidate value with which a minimized vertical error is associated;determine the first measured disparity factor for a longest horizontal baseline of the rectified image array; anddetermine the measured vertical disparity factor for a longest vertical baseline of the rectified image array. 16. The apparatus of claim 10, one or both of the first measured disparity factor or the measured vertical disparity factor comprising a number of pixels. 17. A method, comprising: determining a first measured disparity factor along a first dimension for a rectified image array comprising a plurality of rectified images by performing a first process comprising: iteratively selecting a first candidate value;measuring a first error associated with the selected first candidate value; andselecting as the first measured disparity factor a first candidate value with which a minimized first error is associated; anddetermining a second measured disparity factor along a second dimension for the rectified image array by performing a second process comprising: iteratively selecting a second candidate value;measuring a second error associated with the selected second candidate value; andselecting as the second measured disparity factor a second candidate value with which a minimized second error is associated, the minimized first error comprising a minimum value of a first pixel matching error function, and the minimized second error comprising a minimum value of a second pixel matching error function. 18. The method of claim 17, the first dimension comprising a horizontal dimension extending along a row of the rectified image array, and the method comprising: determining a measured vertical disparity factor for the rectified image array by performing a third process comprising: iteratively selecting a vertical candidate value;measuring a vertical error associated with the selected vertical candidate value; andselecting as the measured vertical disparity factor a vertical candidate value with which a minimized vertical error is associated;determining an implied horizontal disparity factor for the rectified image array based on the measured vertical disparity factor; anddetermining an implied vertical disparity factor for the rectified image array based on the first measured disparity factor. 19. The method of claim 18, comprising: determining a composite horizontal disparity factor for the rectified image array based on the first measured disparity factor and the implied horizontal disparity factor; anddetermining a composite vertical disparity factor for the rectified image array based on the measured vertical disparity factor and the implied vertical disparity factor. 20. The method of claim 19, comprising: determining a horizontal sub-pixel disparity factor and a vertical sub-pixel disparity factor for the rectified image array by performing a fourth process comprising: iteratively selecting a sub-pixel candidate value pair comprising a horizontal sub-pixel candidate value and a vertical sub-pixel candidate value;measuring a joint error associated with the selected sub-pixel candidate value pair;determining as the horizontal sub-pixel disparity factor a horizontal sub-pixel candidate value comprised within a sub-pixel candidate value pair with which a minimized joint error is associated; anddetermining as the vertical sub-pixel disparity factor a vertical sub-pixel candidate value comprised within the sub-pixel candidate value pair with which the minimized joint error is associated. 21. The method of claim 17, the first pixel matching error function and the second pixel matching error function comprising sums of pairwise block matching error functions. 22. The method of claim 17, the first dimension comprising a horizontal dimension extending along a row of the rectified image array, and the method comprising: determining a measured vertical disparity factor for the rectified image array by performing a second process comprising: iteratively selecting a vertical candidate value;measuring a vertical error associated with the selected vertical candidate value; andselecting as the measured vertical disparity factor a vertical candidate value with which a minimized vertical error is associated.determining the first measured disparity factor for a longest horizontal baseline of the rectified image array; anddetermining the measured vertical disparity factor for a longest vertical baseline of the rectified image array. 23. A system, comprising: a processor circuit;a camera array comprising a plurality of cameras; andan imaging management module to: determine a first measured disparity factor along a first dimension for a rectified image array comprising a plurality of rectified images of the plurality of cameras by a first process arranged to: iteratively select a first candidate value;measure a first error associated with the selected first candidate value; andselect as the first measured disparity factor a first candidate value with which a minimized first error is associated; anddetermine a second measured disparity factor along a second dimension for the rectified image array by a second process arranged to: iteratively select a second candidate value;measure a second error associated with the selected second candidate value; andselect as the second measured disparity factor a second candidate value with which a minimized second error is associated, the minimized first error comprising a minimum value of a first pixel matching error function; the minimized second error comprising a minimum value of a second pixel matching error function, the fixed pixel matching error function and the second pixel matching error function comprising sums of pairwise block matching error functions. 24. The system of claim 23, the first dimension comprising a horizontal dimension extending along a row of the rectified image array, and the imaging management module to: determine a measured vertical disparity factor for the rectified image array by a third process arranged to: iteratively select a vertical candidate value;measure a vertical error associated with the selected vertical candidate value; andselect as the measured vertical disparity factor a vertical candidate value with which a minimized vertical error is associated;determine an implied horizontal disparity factor for the rectified image array based on the measured vertical disparity factor; anddetermine an implied vertical disparity factor for the rectified image array based on the first measured disparity factor. 25. The system of claim 24, the imaging management module to: determine a composite horizontal disparity factor for the rectified image array based on the first measured disparity factor and the implied horizontal disparity factor; anddetermine a composite vertical disparity factor for the rectified image array based on the measured vertical disparity factor and the implied vertical disparity factor. 26. The system of claim 25, the imaging management module to: determine a horizontal sub-pixel disparity factor and a vertical sub-pixel disparity factor for the rectified image array by a fourth process arranged to: iteratively select a sub-pixel candidate value pair comprising a horizontal sub-pixel candidate value and a vertical sub-pixel candidate value;measure a joint error associated with the selected sub-pixel candidate value pair;determine as the horizontal sub-pixel disparity factor a horizontal sub-pixel candidate value comprised within a sub-pixel candidate value pair with which a minimized joint error is associated; anddetermine as the vertical sub-pixel disparity factor a vertical sub-pixel candidate value comprised within the sub-pixel candidate value pair with which the minimized joint error is associated.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (16)
Barratt, Dean; Hu, Yipeng, Apparatus and method for registering two medical images.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.