公开(公告)号：US20130208977A1

公开(公告)日：2013-08-15

申请号：US13668188

申请日：2012-11-02

Applicant: NEC Laboratories America, Inc.

Inventor： Yangqing Jia ,  Chang Huang

IPC: G06K9/62

CPC classification number: G06K9/6256 ,  G06K2009/4695

Abstract: Systems and methods are disclosed for image classification by receiving an overcomplete set of spatial regions, jointly optimizing the classifier and the pooling region for each pooled feature; and performing incremental feature selection and retraining using a grafting process to efficiently train the classifier.

Abstract translation: 公开了用于图像分类的系统和方法，其通过接收空间区域的不完整集合，联合优化每个合并特征的分类器和汇集区域; 并使用移植过程执行增量特征选择和再培训，以有效地训练分类器。

公开(公告)号：US08873844B2

公开(公告)日：2014-10-28

申请号：US13682780

申请日：2012-11-21

Applicant: NEC Laboratories America, Inc.

Inventor： Chang Huang ,  Shenghuo Zhu ,  Kai Yu

IPC: G06K9/62

CPC classification number: G06K9/6256 ,  G06K9/6232

Abstract: Systems and methods for metric learning include iteratively determining feature groups of images based on its derivative norm. Corresponding metrics of the feature groups are learned by gradient descent based on an expected loss. The corresponding metrics are combined to provide an intermediate metric matrix as a sparse representation of the images. A loss function of all metric parameters corresponding to features of the intermediate metric matrix are optimized, using a processor, to learn a final metric matrix. Eigenvalues of the final metric matrix are projected onto a simplex.

Abstract translation: 度量学习的系统和方法包括基于其导数规范迭代确定图像的特征组。 通过基于预期损失的梯度下降来学习特征组的相应度量。 相应的度量被组合以提供作为图像的稀疏表示的中间度量矩阵。 使用处理器来优化对应于中间度量矩阵的特征的所有度量参数的损失函数来学习最终的度量矩阵。 最终公制矩阵的特征值被投影到单纯形上。

公开(公告)号：US08781218B2

公开(公告)日：2014-07-15

申请号：US13668188

申请日：2012-11-02

Applicant: NEC Laboratories America, Inc.

Inventor： Yangqing Jia ,  Chang Huang

IPC: G06K9/62 ,  G06K9/46

CPC classification number: G06K9/6256 ,  G06K2009/4695

Abstract: Systems and methods are disclosed for image classification by receiving an overcomplete set of spatial regions, jointly optimizing the classifier and the pooling region for each pooled feature; and performing incremental feature selection and retraining using a grafting process to efficiently train the classifier.

Abstract translation: 公开了用于图像分类的系统和方法，其通过接收空间区域的不完整集合，联合优化每个合并特征的分类器和汇集区域; 并使用移植过程执行增量特征选择和再培训，以有效地训练分类器。

公开(公告)号：US20130129202A1

公开(公告)日：2013-05-23

申请号：US13682780

申请日：2012-11-21

Applicant: NEC Laboratories America, Inc.

Inventor： Chang Huang ,  Shenghuo Zhu ,  Kai Yu

IPC: G06K9/62

CPC classification number: G06K9/6256 ,  G06K9/6232

Abstract: Systems and methods for metric learning include iteratively determining feature groups of images based on its derivative norm. Corresponding metrics of the feature groups are learned by gradient descent based on an expected loss. The corresponding metrics are combined to provide an intermediate metric matrix as a sparse representation of the images. A loss function of all metric parameters corresponding to features of the intermediate metric matrix are optimized, using a processor, to learn a final metric matrix. Eigenvalues of the final metric matrix are projected onto a simplex.

Abstract translation: 度量学习的系统和方法包括基于其导数规范迭代确定图像的特征组。 通过基于预期损失的梯度下降来学习特征组的相应度量。 相应的度量被组合以提供作为图像的稀疏表示的中间度量矩阵。 使用处理器来优化对应于中间度量矩阵的特征的所有度量参数的损失函数来学习最终的度量矩阵。 最终公制矩阵的特征值被投影到单纯形上。