US8031914B2
Expired / Lapsed
US
⏱️ 0.0 years remaining
Face-based image clustering
Assignee: Hewlett Packard Development Co LP
•
Filing: 2006-10-11
•
Granted: 2011-10-04
•
Expires: 2026-10-11
Abstract
Abstract Face-based image clustering systems and methods are described. In one aspect, face regions are detected in images. At least one respective parameter value is extracted from each of the face regions. Ones of the face regions associated with parameter values satisfying a cluster seed predicate are classified as cluster seed face regions. The cluster seed face regions are clustered into one or more clusters. A respective face model is built for each of the clusters. The face models are stored. In another aspect, face regions are detected in images. At least one respective parameter value is extracted from each of the face regions. The face regions are ranked based on the extracted parameter values. The face regions are clustered in rank order into one or more clusters. Representations of ones of the clusters are rendered on a display.
Inventors
- 👤 Tong Zhang
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-09 10:06
CPC Classifications
G; G06; G06F; G06F16/00; G06F16/50; G06F16/58; G06F16/583; G06F16/5854; G06F18/00; G06F18/20; G06F18/21; G06F18/211; G06F18/2113; G06V; G06V10/00; G06V10/70; G06V10/77; G06V10/771; G06V10/98; G06V10/993; G06V40/00; G06V40/10; G06V40/16; G06V40/161
Claims Analysis
Total Claims: 25
Claim 1
Independent Claim
67 words
A method, comprising: detecting face regions in images; extracting at least one respective parameter value from each of the face regions; classifying as cluster seed face regions ones of the face regions associated with parameter values satisfying a cluster seed predicate; clustering the cluster seed face regions into one or more clusters; building a respective face model for each of the clusters; and storing the face models.
Claim 2
Independent Claim
67 words
A method, comprising: detecting face regions in images; extracting at least one respective parameter value from each of the face regions; classifying as cluster seed face regions ones of the face regions associated with parameter values satisfying a cluster seed predicate; clustering the cluster seed face regions into one or more clusters; building a respective face model for each of the clusters; and storing the face models.
Claim 3
Dependent Claim
60 words
The method of claim 1 , wherein the extracting comprises determining for each of the face regions respective values for one or more of the following parameters: a measure of sharpness of the face region; a measure of contrast of the face region; a measure of saturation of the face region; and a measure of exposure of the face region.
Claim 4
Dependent Claim
21 words
The method of claim 1 , wherein the extracting comprises determining respective size parameter values specifying sizes of the face regions.
Claim 5
Dependent Claim
37 words
The method of claim 3 , wherein the determining comprises determining the respective size parameter values specifying the sizes of the face regions in relation to sizes of corresponding ones of the images containing the face regions.
Claim 6
Dependent Claim
49 words
The method of claim 1 , wherein the classifying comprises classifying as non-cluster-seed face regions each of the face regions associated with at least one of: a brightness parameter value that fails to meet a brightness predicate; and a sharpness parameter value that fails to meet a sharpness predicate.
Claim 7
Dependent Claim
38 words
The method of claim 5 , wherein the clustering comprises assigning ones of the non-cluster-seed face regions to respective ones of the clusters based on measures of nearness between the non-cluster-seed face regions and the corresponding face models.
Claim 8
Dependent Claim
121 words
The method of claim 1 , wherein the clustering comprises: (a) initiating a cluster with one of the cluster seed face regions; (b) determining a respective measure of nearness between each initiated cluster and a next one of the cluster seed face regions; (c) assigning the next cluster seed face region to the nearest one of the initiated clusters in response to a determination that the corresponding nearness measure satisfies a first nearness predicate; (d) initiating a new cluster with the next cluster seed face region in response to a determination that the corresponding nearness measure satisfies a second nearness predicate different from the first nearness predicate; and (e) repeating (b)-(d) for all successive ones of the cluster seed face regions.
Claim 9
Dependent Claim
47 words
The method of claim 7 , wherein, before (a), the clustering comprises ranking the cluster seed face regions based on the values extracted for at least one parameter, and (b) comprises selecting the next cluster seed face region in rank order from highest rank to lowest rank.
Claim 10
Dependent Claim
54 words
The method of claim 7 , wherein (b) comprises the building of a respective face model for each initiated cluster based on features extracted from its constituent cluster seed face regions and determining a respective measure of nearness between features extracted from the next cluster seed face region and each of the face models.
Claim 11
Dependent Claim
47 words
The method of claim 7 , wherein, after (e), the clustering comprises (f) assigning to the nearest ones of the initiated clusters unassigned ones of the cluster seed face regions associated with nearness measures satisfying a third nearness predicate different from the first and second nearness predicates.
Claim 12
Dependent Claim
26 words
The method of claim 10 , wherein, after (f), repeating (a)-(e) for ones of the cluster seed face regions unassigned to any of the initiated clusters.
Claim 13
Dependent Claim
27 words
The method of claim 1 , wherein the clustering comprises automatically merging clusters containing respective numbers of constituent face regions that fail to meet a count threshold.
Claim 14
Dependent Claim
25 words
The method of claim 1 , wherein the clustering comprises preventing clusters from containing more than one constituent face region from each of the images.
Claim 15
Dependent Claim
41 words
The method of claim 1 , further comprising rendering representations of ones of the clusters on a display, modifying ones of the rendered clusters in response to user input, and rebuilding a respective face model for each of the modified clusters.
Claim 16
Dependent Claim
35 words
The method of claim 1 , wherein the classifying comprises classifying as non-cluster-seed face regions each of the face regions associated with one or more size parameter values that fail to meet a size predicate.
Claim 17
Independent Claim
104 words
An apparatus, comprising; a computer-readable medium storing computer-readable instructions; and a processor coupled to the computer-readable medium, operable to execute the instructions, and based at least in part on the execution of the instructions operable to perform operations comprising detecting face regions in images, extracting at least one respective parameter value from each of the face regions, classifying as cluster seed face regions ones of the face regions associated with parameter values satisfying a cluster seed predicate, clustering the cluster seed face regions into one or more clusters, and building a respective face model for each of the clusters, and store the face models.
Claim 18
Independent Claim
104 words
An apparatus, comprising; a computer-readable medium storing computer-readable instructions; and a processor coupled to the computer-readable medium, operable to execute the instructions, and based at least in part on the execution of the instructions operable to perform operations comprising detecting face regions in images, extracting at least one respective parameter value from each of the face regions, classifying as cluster seed face regions ones of the face regions associated with parameter values satisfying a cluster seed predicate, clustering the cluster seed face regions into one or more clusters, and building a respective face model for each of the clusters, and store the face models.
Claim 19
Dependent Claim
51 words
The apparatus of claim 16 , wherein the processor is operable to classify as non-cluster-seed face regions each of the face regions associated with at least one of: a brightness parameter value that fails to meet a brightness predicate; and a sharpness parameter value that fails to meet a sharpness predicate.
Claim 20
Dependent Claim
40 words
The apparatus of claim 17 , wherein the processor is operable to assign ones of the non-cluster-seed face regions to respective ones of the clusters based on measures of nearness between the non-cluster-seed face regions and the corresponding face models.
Claim 21
Dependent Claim
126 words
The apparatus of claim 16 , wherein the processor is operable to perform operations comprising: (a) initiating a cluster with one of the cluster seed face regions; (b) determining a respective measure of nearness between each initiated cluster and a next one of the cluster seed face regions; (c) assigning the next cluster seed face region to the nearest one of the initiated clusters in response to a determination that the corresponding nearness measure satisfies a first nearness predicate; (d) initiating a new cluster with the next cluster seed face region in response to a determination that the corresponding nearness measure satisfies a second nearness predicate different from the first nearness predicate; and (e) repeating (b)-(d) for all successive ones of the cluster seed face regions.
Claim 22
Dependent Claim
35 words
The apparatus of claim 16 , wherein the classifying comprises classifying as non-cluster-seed face regions each of the face regions associated with one or more size parameter values that fail to meet a size predicate.
Claim 23
Independent Claim
54 words
A method, comprising: detecting face regions in images; extracting at least one respective parameter value from each of the face regions; ranking the face regions based on the extracted parameter values; clustering the face regions in rank order into one or more clusters; and rendering representations of ones of the clusters on a display.
Claim 24
Independent Claim
54 words
A method, comprising: detecting face regions in images; extracting at least one respective parameter value from each of the face regions; ranking the face regions based on the extracted parameter values; clustering the face regions in rank order into one or more clusters; and rendering representations of ones of the clusters on a display.
Claim 25
Dependent Claim
42 words
The method of claim 21 , further comprising building a respective face model for each of the clusters, modifying one or more of the clusters in response to user input, and rebuilding a respective face model for each of the modified clusters.
References Cited (Backward Citations)
USRE36041E
Cited
US5418895A
Cited
US5550928A
Cited
US20030179911A1
Cited
US20010043727A1
Cited
US6282317B1
Cited
US6535636B1
Cited
US20020150280A1
Cited
US20030048950A1
Cited
US20030068100A1
Cited
US20040197013A1
Cited
US20030210808A1
Cited
US20040022442A1
Cited
WO2004044823A2
Cited
US20050105806A1
Cited
US20050105780A1
Cited
US20050265581A1
Cited
US20060204058A1
Cited
US20060140455A1
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2014-10-03 | 2015-04-04 | 2015-10-04 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2018-10-03 | 2019-04-04 | 2019-10-04 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2022-10-03 | 2023-04-04 | 2023-10-04 | $7700.00 | Window Closed / Lapsed |