Clustering large database of images using multilevel clustering approach for optimized face recognition process
원문보기
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
G06K-009/00
G06K-009/62
출원번호
US-0228706
(2016-08-04)
등록번호
US-9904844
(2018-02-27)
발명자
/ 주소
Asati, Somnath
Eshwar, Bhavani K.
Naganna, Soma Shekar
Seth, Abhishek
Tomar, Vishal
출원인 / 주소
International Business Machines Corporation
대리인 / 주소
North Shore Patents, P.C.
인용정보
피인용 횟수 :
0인용 특허 :
15
초록▼
In multilevel clustering for a face recognition process, the first stage clustering is performed on each computing node, using the first x vector coefficients. From the resulting k clusters created in the first stage, a limited number of clusters are selected on which the second stage clustering is
In multilevel clustering for a face recognition process, the first stage clustering is performed on each computing node, using the first x vector coefficients. From the resulting k clusters created in the first stage, a limited number of clusters are selected on which the second stage clustering is performed, using the next y vector coefficients. The search for a matching image is then limited to these selected clusters. Computational costs are reduced at the first stage clustering by using just the first x vector coefficients. Computational costs for the second stage clustering are also reduced by performing the second stage only with the limited number of clusters on a limited number of computing nodes. In this manner, the overall computational costs in the face recognition process is significantly reduced while maintaining a desired level of accuracy.
대표청구항▼
1. A computer program product for multilevel clustering for a face recognition process, the computer program product comprising one or more computer readable storage media having program instructions embodied therewith, the program instructions executable by one or more processors to: perform a firs
1. A computer program product for multilevel clustering for a face recognition process, the computer program product comprising one or more computer readable storage media having program instructions embodied therewith, the program instructions executable by one or more processors to: perform a first stage of clustering of an image data set by the one or more computer systems, the image data set comprising a plurality of image vectors representing a plurality of facial images, comprising: choose k vectors in the image data set as a set of k clusters; andassign each of the remaining image vectors in the image data set to any of the k clusters using first x vector coefficients;calculate a first distance between a query image vector representing a query image and each of the k clusters using the first x vector coefficients;select at least a first cluster and a second cluster from the k clusters for which the first distance is minimum;perform a second stage of clustering with the first cluster and the second cluster by the one or more computer systems, comprising: choose first m image vectors in the first cluster as a set of first m sub-clusters, and assign each of the remaining image vectors in the first cluster to any of the m sub-clusters using next y vector coefficients; andchoose first m image vectors in the second cluster as a set of second m sub-clusters and assign each of the remaining image vectors in the second cluster to any of the second m sub-clusters using the next y vector coefficients;calculate a second distance between the query image vector and the first and second m sub-clusters using the next y vector coefficients;select a first sub-cluster in the first m sub-clusters and a second sub-cluster in the second m sub-clusters for which the second distance is minimum; andselect a given image vector from either the first sub-cluster or the second sub-cluster as matching the query image vector. 2. The computer program product of claim 1, wherein the assigning of each of the remaining image vectors in the image data set to any of the k clusters using the first x vector coefficients comprises: calculate a third distance between a given image vector in the image data set and each of the k clusters using the first x vector coefficients; andassign the given vector to a given cluster of the k clusters for which the third distance is minimum. 3. The computer program product of claim 1, wherein the first stage of clustering is performed at each of a plurality of computer systems, wherein the selecting of at least the first cluster and the second cluster from the k clusters for which the first distance is minimum comprises: at each of the plurality of computer systems, select a given cluster from the k clusters for which the first distance is minimum;calculate a third distance between the query image vector and each of the selected given clusters; andselect at least the first cluster at a first computer system and the second cluster at the second computer system for which the third distance is minimum,wherein the second stage of clustering is performed with the first and second clusters at the first and second computer systems. 4. The computer program product of claim 1, wherein the assigning of each of the remaining image vectors in the first cluster to any of the m sub-clusters using the next y vector coefficients comprises: calculate a third distance between a given vector in the first cluster and each of the m sub-clusters using the next y vector coefficients; andassign the given vector in the first cluster to a given sub-cluster of the m sub-clusters for which the third distance is minimum. 5. The computer program product of claim 1, wherein the selecting of the given image vector from either the first sub-cluster or the second sub-cluster as matching the query image vector comprises: calculate a third distance between the query image vector and each of the image vectors in the first sub-cluster and selecting a first nearest image vector in the first sub-cluster for which the third distance is minimum; andcalculate a fourth distance between the query image vector and each of the image vectors in the second sub-cluster and selecting a second nearest image vector in the second sub-cluster for which the fourth distance is minimum. 6. The computer program product of claim 5, wherein the selecting of the given image vector from either the first sub-cluster or the second sub-cluster as matching the query image vector further comprises: calculate a fifth distance between the query image vector and the first nearest image vector;calculate a sixth distance between the query image vector and the second nearest image vector; andselect either the first nearest image vector or the second nearest image vector for which either the fifth distance or the sixth distance is minimum. 7. The computer program product of claim 1, further comprising: determine whether the given image vector is within a predetermined similarity threshold; andbased on determining that the given image vector is within the predetermined similarity threshold, output an image represented by the given image vector as matching the query image. 8. A system, comprising: one or more processors; andone or more computer readable storage media having program instructions embodied therewith, the program instructions executable by the one or more processors to: perform a first stage of clustering of an image data set by the one or more computer systems, the image data set comprising a plurality of image vectors representing a plurality of facial images, comprising: choose k vectors in the image data set as a set of k clusters; andassign each of the remaining image vectors in the image data set to any of the k clusters using first x vector coefficients;calculate a first distance between a query image vector representing a query image and each of the k clusters using the first x vector coefficients;select at least a first cluster and a second cluster from the k clusters for which the first distance is minimum;perform a second stage of clustering with the first cluster and the second cluster by the one or more computer systems, comprising: choose first m image vectors in the first cluster as a set of first m sub-clusters, and assign each of the remaining image vectors in the first cluster to any of the m sub-clusters using next y vector coefficients; andchoose first m image vectors in the second cluster as a set of second m sub-clusters and assign each of the remaining image vectors in the second cluster to any of the second m sub-clusters using the next y vector coefficients;calculate a second distance between the query image vector and the first and second m sub-clusters using the next y vector coefficients;select a first sub-cluster in the first m sub-clusters and a second sub-cluster in the second m sub-clusters for which the second distance is minimum; andselect a given image vector from either the first sub-cluster or the second sub-cluster as matching the query image vector. 9. The system of claim 8, wherein the assigning of each of the remaining image vectors in the image data set to any of the k clusters using the first x vector coefficients comprises: calculate a third distance between a given image vector in the image data set and each of the k clusters using the first x vector coefficients; andassign the given vector to a given cluster of the k clusters for which the third distance is minimum. 10. The system of claim 8, wherein the first stage of clustering is performed at each of a plurality of computer systems, wherein the selecting of at least the first cluster and the second cluster from the k clusters for which the first distance is minimum comprises: at each of the plurality of computer systems, select a given cluster from the k clusters for which the first distance is minimum;calculate a third distance between the query image vector and each of the selected given clusters; andselect at least the first cluster at a first computer system and the second cluster at the second computer system for which the third distance is minimum,wherein the second stage of clustering is performed with the first and second clusters at the first and second computer systems. 11. The system of claim 8, wherein the assigning of each of the remaining image vectors in the first cluster to any of the m sub-clusters using the next y vector coefficients comprises: calculate a third distance between a given vector in the first cluster and each of the m sub-clusters using the next y vector coefficients; andassign the given vector in the first cluster to a given sub-cluster of the m sub-clusters for which the third distance is minimum. 12. The system of claim 8, wherein the selecting of the given image vector from either the first sub-cluster or the second sub-cluster as matching the query image vector comprises: calculate a third distance between the query image vector and each of the image vectors in the first sub-cluster and selecting a first nearest image vector in the first sub-cluster for which the third distance is minimum; andcalculate a fourth distance between the query image vector and each of the image vectors in the second sub-cluster and selecting a second nearest image vector in the second sub-cluster for which the fourth distance is minimum. 13. The system of claim 12, wherein the selecting of the given image vector from either the first sub-cluster or the second sub-cluster as matching the query image vector further comprises: calculate a fifth distance between the query image vector and the first nearest image vector;calculate a sixth distance between the query image vector and the second nearest image vector; andselect either the first nearest image vector or the second nearest image vector for which either the fifth distance or the sixth distance is minimum.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (15)
Luo, Jiebo; Gallagher, Andrew C., Classifying image regions based on picture location.
Boucheron, Laura E., Combinational pixel-by-pixel and object-level classifying, segmenting, and agglomerating in performing quantitative image analysis that distinguishes between healthy non-cancerous and cancerous cell nuclei and delineates nuclear, cytoplasm, and stromal material objects from stained biological tissue materials.
Stavely,Donald J.; Hall,Kenneth Jay; Battles,Amy E.; Barrios,Sarah Jean; Sobol,Robert E., Digital imaging method and apparatus using eye-tracking control.
Johnson, Steve Gail; Clark, Matthew Cameron; Ecker, Stephen; Johnson, Kyle Brook; Kuether, Richard Craig, System and method for using a video monitoring system to prevent and manage decubitus ulcers in patients.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.