| EUC003US |
Australia, Canada, China (Peoples Republic), European Patent Convention, Hong Kong, India, Japan, S. Korea |
Patent Issued (US, China,
Australia)
US No. 7158680
Published
|
First filing - describes process that initially includes the high level steps of unsupervised object modeling though feature/object detection and tracking, foreground/background segmentation, and spatial normalization.
|
10 allowed claims. |
| EUC003USCON |
|
Patent Issued (US)
No. 7424157
Published |
Continuation of EUC003US, claims same priority date as first filing but claims expand coverage beyond the 10 original claims. |
19 allowed claims. |
| EUC004US |
Australia, China (Peoples Republic), European Patent Convention, Hong Kong, India, Japan, S. Korea |
Patent Issued (US, Australia)
No. 7426285
Published |
Provides additional benefit by combining the decomposed data with conventional CODEC and fine grained pel/subpel analysis, 3D normalization, progressive geometric normalization, subspace classification, and hybrid spatial compression techniques. |
16 allowed claims. |
| EUC005US |
Australia, China (Peoples Republic), European Patent Convention, Hong Kong, India, Japan, S. Korea |
Patent Issued (US, China,
Australia)
No. 7436981
Published
|
The innovation is expanded in terms of the re-sampling of pel data used to create the representation of an object in the video. Techniques primarily included are the application of a regular spatial mesh to the localized appearance of the object, use of 2D affine models to locally interpolate pel values, and factoring global object motion from local object deformation. |
17 allowed claims. |
| EUC007US |
Australia, China (Peoples Republic), European Patent Convention, Hong Kong, India, Japan, S. Korea |
Patent Issued (US, Australia)
No. 7457435
Published |
Refines background resolution and adds mesh-based registration and motion estimation. Models correspondences to generate contour models that are used to segment the object from non-object background in the video frame sequence. |
14 allowed claims. |
| EUC009US |
Australia, China (Peoples Republic), European Patent Convention, Hong Kong, India, Japan (2 filings), S. Korea |
Patent Issued (US, China, Japan)
No. 7457472
Published |
The previously disclosed innovative material is expanded to cover at least the explicit implementation embodied in the Phase II prototype. Highlights include greatly expanding the PCA processing into a "mixture of experts" architecture along with high dimensional subspace partitioning for both the appearance and deformation manifolds. |
10 allowed claims. |
| EUC011US |
Canada, China (Peoples Republic), European Patent Convention, Hong Kong, Japan, Taiwan |
Patent Issued (US)
No. 7508990
Published |
The progressive computational environment is disclosed; an environment that allows for the progressive analysis of the video stream to determine the amount of modeling that can be performed given the complexity of the video's data and the computational resource constraints provided. |
21 allowed claims. |
| EUC012US |
Canada, China (Peoples Republic), European Patent Convention, Hong Kong, Japan, Taiwan |
Published |
The hybrid utilization of EuclidVision inside of a conventional encoder architecture is disclosed. This is the first of three patents families that were filed concurrently. These three patent families represent the technology offering in terms of its application in the consumer marketplace. |
23 pending claims. |
| EUC013US |
Canada, China (Peoples Republic), European Patent Convention, Hong Kong, Japan, Taiwan |
Published |
Specific disclosure of an embodiment of single object EuclidVision for video conferencing where a predetermined 3D face/head appearance/structural model is calibrated to the face in the video and then used to model the appearance of the face at various points in the video. |
8 pending claims. |
| EUC014US |
US, Canada, European Patent Office, China, Japan, Taiwan, Hong Kong |
Published |
Applies EuclidVision techniques to an Object Archival System business application patent. Specifically provides for the ability for image based search and retrieval of video. Supervised and unsupervised object identification are disclosed along with a multitude of service network topologies. |
25 pending claims. |
| EUC015 |
Patent Cooperation Treaty, Taiwan |
Published at PCT |
Describes modeling features within a conventional video encoding framework. Motion compensated prediction is augmented with feature tracking information that predicts past feature positions while also predicting the appearance of the feature in the current frame. |
35 pending claims. |