公开(公告)号：US20160307061A1

公开(公告)日：2016-10-20

申请号：US14688255

申请日：2015-04-16

Applicant: Xerox Corporation

Inventor： Orhan Bulan ,  Claude Fillion ,  Aaron M. Burry ,  Vladimir Kozitsky

IPC: G06K9/32 ,  G06K9/62 ,  G06K9/00 ,  G06K9/18

CPC classification number: G06K9/6256 ,  G06K2209/01 ,  G06K2209/15

Abstract: Methods and systems for bootstrapping an OCR engine for license plate recognition. One or more OCR engines can be trained utilizing purely synthetically generated characters. A subset of classifiers, which require augmentation with real examples, along how many real examples are required for each, can be identified. The OCR engine can then be deployed to the field with constraints on automation based on this analysis to operate in a “bootstrapping” period wherein some characters are automatically recognized while others are sent for human review. The previously determined number of real examples required for augmenting the subset of classifiers can be collected. Each subset of identified classifiers can then be retrained as the number of real examples required becomes available.

Abstract translation: 引导OCR引擎进行车牌识别的方法和系统。 可以使用纯合成生成的字符来训练一个或多个OCR引擎。 可以识别分类器的一个子集，它们需要通过实例增加，每个需要多少个实例。 然后，可以将OCR引擎部署到具有基于该分析的自动化约束的现场，以在“自举”期间操作，其中一些字符被自动识别，而其他人被发送供人审查。 可以收集先前确定的用于增加分类器子集所需的实例的数量。 随着所需实际数量的可用数量的增加，识别的分类器的每个子集可以被重新训练。

公开(公告)号：US20140348392A1

公开(公告)日：2014-11-27

申请号：US13899742

申请日：2013-05-22

Applicant: XEROX CORPORATION

Inventor： Aaron Michael Burry ,  Vladimir Kozitsky

IPC: G06K9/32

CPC classification number: G06K9/325 ,  G06K2209/15

Abstract: Methods and systems for automatically determining the issuing state of a license plate. An image of a license plate acquired by an ALPR engine can be processed via one or more OCR engines such that each OCR engine among the OCR engines is tuned to a particular state. Confidence data output from the OCR engine(s) can be analyzed (among other factors) to estimate the issuing state associated with the license plate. Multiple observations related to the issuing state can be merged to derive an overall conclusion and assign an associated confidence value with respect to the confidence data and determine a likely issuing state associated with the license plate.

Abstract translation: 自动确定车牌发卡状态的方法和系统。 可以通过一个或多个OCR引擎来处理由ALPR引擎获取的车牌的图像，使得OCR引擎中的每个OCR引擎被调谐到特定状态。 可以分析从OCR引擎输出的置信度数据（以及其他因素）来估计与车牌相关的发卡状态。 可以合并与发布状态有关的多个观察结果，以获得总体结论，并且相对于置信度数据分配相关的置信度值，并确定与该牌照相关联的可能发布状态。

公开(公告)号：US20140301606A1

公开(公告)日：2014-10-09

申请号：US13857657

申请日：2013-04-05

Applicant: XEROX CORPORATION

Inventor： Peter Paul ,  Vladimir Kozitsky ,  Aaron Michael Burry ,  Martin Edward Hoover ,  Mark Cantelli

IPC: G06K9/18

CPC classification number: G06K9/2027 ,  G06K9/3258 ,  G06K2209/15

Abstract: Methods, systems, and processor-readable media for data augmentation utilized in an automatic license plate recognition engine. A machine-readable code can be associated with an automatic license plate recognition engine. The machine-readable code can be configured to define parameters that drive processing within the automatic license plate recognition engine to produce recognition results thereof and enhance a machine readability of a license plate recognized and analyzed via the automatic license plate recognition engine.

Abstract translation: 用于在自动车牌识别引擎中使用的用于数据增加的方法，系统和处理器可读介质。 机器可读代码可以与自动车牌识别引擎相关联。 机器可读代码可以被配置为定义驱动自动车牌识别引擎内的处理以产生其识别结果的参数，并提高通过自动车牌识别引擎识别和分析的牌照的机器可读性。

公开(公告)号：US20140079315A1

公开(公告)日：2014-03-20

申请号：US13746412

申请日：2013-01-22

Applicant: XEROX CORPORATION

Inventor： Vladimir Kozitsky ,  Aaron Michael Burry ,  Peter Paul

IPC: G06K9/62

CPC classification number: A47C7/14 ,  B60N2/02 ,  B60N2/1803 ,  B60N2/1835 ,  B60N2/1864 ,  B60N2002/0236 ,  G06K9/3258 ,  G06K9/623 ,  G06K9/6267 ,  G06K2009/4666 ,  G06K2209/01 ,  G06K2209/15

Abstract: Methods and systems for reducing the required footprint of SNoW-based classifiers via optimization of classifier features. A compression technique involves two training cycles. The first cycle proceeds normally and the classifier weights from this cycle are used to rank the Successive Mean Quantization Transform (SMQT) features using several criteria. The top N (out of 512 features) are then chosen and the training cycle is repeated using only the top N features. It has been found that OCR accuracy is maintained using only 60 out of 512 features leading to an 88% reduction in RAM utilization at runtime. This coupled with a packing of the weights from doubles to single byte integers added a further 8× reduction in RAM footprint or a reduction of 68× over the baseline SNoW method.

Abstract translation: 通过优化分类器功能来减少基于SNoW的分类器所需占用面积的方法和系统。 压缩技术涉及两个训练周期。 第一个周期正常进行，并且使用来自该周期的分类器权重使用几个标准对连续平均量化变换（SMQT）特征进行排序。 然后选择顶部N（512个特征之​​一），并且仅使用前N个特征重复训练周期。 已经发现，使用仅512个特征中的60个来维持OCR精度，导致运行时RAM利用率降低88％。 这加上权重从双精度到单字节整数的加载，增加了RAM占用的8倍减少或基线SNoW方法减少了68倍。

公开(公告)号：US10026004B2

公开(公告)日：2018-07-17

申请号：US15205146

申请日：2016-07-08

Applicant: Xerox Corporation

Inventor： Howard Mizes ,  Vladimir Kozitsky

IPC: G06K9/34 ,  G06T5/00 ,  G06K9/00 ,  G06T7/00

Abstract: A method, system, and apparatus for license plate relighting comprises collecting an image of a license plate, performing license plate recognition on the image of the license plate; calculating a confidence metric for the license plate recognition; and performing a shadow detection and relighting method if the confidence metric is below a predetermined threshold, comprising identifying a shaded region of said license plate, determining if the shaded region is actually shaded, and relighting the actually shaded region.

公开(公告)号：US20170262723A1

公开(公告)日：2017-09-14

申请号：US15064167

申请日：2016-03-08

Applicant: Xerox Corporation

Inventor： Vladimir Kozitsky ,  Matthew Shreve ,  Orhan Bulan

IPC: G06K9/32 ,  G06K9/62

CPC classification number: G06K9/325 ,  G06K9/3216 ,  G06K9/4628 ,  G06K9/6267 ,  G06K9/6857 ,  G06K2209/15

Abstract: Methods, systems, and processor-readable media for the detection and classification of license plates. In an example embodiment, an image of a vehicle can be captured with an image-capturing unit. A license plate region can then be located in the captured image of the vehicle by extracting a set of candidate regions from the image utilizing a weak classifier. A set of candidate regions can be ranked utilizing a secondary strong classifier. The captured image can then be classified according to a confidence driven classification based on classification criteria determined by the weak classifier and the secondary strong classifier.

公开(公告)号：US20170161581A1

公开(公告)日：2017-06-08

申请号：US14960627

申请日：2015-12-07

Applicant: Xerox Corporation

Inventor： Vladimir Kozitsky ,  Martin S. Maltz

IPC: G06K9/32 ,  G06T7/40 ,  G06T7/00

CPC classification number: G06K9/3258 ,  G06K2209/01 ,  G06K2209/15 ,  G06T7/11 ,  G06T7/90 ,  G06T2207/10024 ,  G06T2207/30252

Abstract: Methods and systems for exploiting color for object recognition. A color gradient for each pixel in a gradient image of an object can be calculated. The gradient image can be binarized to produce an image having white walls around characters and other structures. A connected component analysis operation can be performed with respect to black regions in the image to determine bounding boxes for characters and other pictorial elements in the image and thereafter identify character candidates from the image utilizing character metrics. Non-character colors can then be eliminated from the image utilizing an outlier rejection.

公开(公告)号：US09536315B2

公开(公告)日：2017-01-03

申请号：US14658713

申请日：2015-03-16

Applicant: Xerox Corporation

Inventor： Orhan Bulan ,  Vladimir Kozitsky ,  Aaron M. Burry

IPC: G06K9/00 ,  G06T7/00 ,  G06K9/62 ,  G06K9/32

CPC classification number: G06T7/0079 ,  G06K9/3258 ,  G06K9/6256 ,  G06K9/6277 ,  G06K2209/01

Abstract: Methods and systems for recognizing a license plate character. Synthetic license plate character images are generated for a target jurisdiction. A limited set of license plate images can be captured for a target jurisdiction utilizing an image-capturing unit. The license plate images are then segmented into license plate character images for the target jurisdiction. The license plate character images collected for the target jurisdiction can be manually labeled. A domain adaptation technique can be utilized to reduce the divergence between synthetically generated and manually labeled target jurisdiction image sets. Additionally, OCR classifiers are trained utilizing the images after the domain adaptation method has been applied. One or more input license plate character images can then be received from the target jurisdiction. Finally, the trained OCR classifier can be employed to determine the most likely labeling for the character image and a confidence associated with the label.

Abstract translation: 识别车牌字符的方法和系统。 为目标管辖区生成综合车牌字符图像。 使用图像捕获单元，可以针对目标管辖区捕获有限的车牌图像。 然后将车牌图像分割为目标司法管辖区的车牌字符图像。 为目标管辖区收集的车牌字符图像可以手动标记。 可以利用域适配技术来减少合成生成的和手动标记的目标管辖图像集之间的差异。 另外，在应用域适配方法之后，使用图像对OCR分类器进行训练。 然后可以从目标管辖区接收一个或多个输入牌照字符图像。 最后，训练有素的OCR分类器可用于确定人物图像最可能的标记和与标签相关的置信度。

公开(公告)号：US20160148076A1

公开(公告)日：2016-05-26

申请号：US14561512

申请日：2014-12-05

Applicant: Xerox Corporation

Inventor： Vladimir Kozitsky ,  Matthew Shreve ,  Aaron M. Burry

IPC: G06K9/62 ,  G06N99/00

CPC classification number: G06K9/627 ,  G06K9/00771 ,  G06K9/325 ,  G06K2209/15 ,  G06N99/005 ,  G08G1/0175

Abstract: Systems and methods for automating an image rejection process. Features including texture, spatial structure, and image quality characteristics can be extracted from one or more images to train a classifier. Features can be calculated with respect to a test image for submission of the features to the classifier, given an operating point corresponding to a desired false positive rate. One or more inputs can be generated from the classifier as a confidence value corresponding to a likelihood of, for example: a license plate being absent in the image, the license plate being unreadable, or the license plate being obstructed. The confidence value can be compared against a threshold to determine if the image(s) should be removed from a human review pipeline, thereby reducing images requiring human review.

Abstract translation: 用于自动化镜像抑制过程的系统和方法。 可以从一个或多个图像中提取包括纹理，空间结构和图像质量特征的特征来训练分类器。 给定对应于所需假阳性率的操作点，可以相对于用于将特征提交给分类器的测试图像来计算特征。 可以从分类器产生一个或多个输入作为对应于例如图像中不存在牌照，牌照不可读或牌照妨碍的可能性的置信度值。 可以将置信度值与阈值进行比较，以确定图像是否应从人类审查管道中移除，从而减少需要人工审查的图像。

公开(公告)号：US09082037B2

公开(公告)日：2015-07-14

申请号：US13899742

申请日：2013-05-22

Applicant: Xerox Corporation

Inventor： Aaron Michael Burry ,  Vladimir Kozitsky

IPC: G06K9/00 ,  G06K9/32

CPC classification number: G06K9/325 ,  G06K2209/15

Abstract: Methods and systems for automatically determining the issuing state of a license plate. An image of a license plate acquired by an ALPR engine can be processed via one or more OCR engines such that each OCR engine among the OCR engines is tuned to a particular state. Confidence data output from the OCR engine(s) can be analyzed (among other factors) to estimate the issuing state associated with the license plate. Multiple observations related to the issuing state can be merged to derive an overall conclusion and assign an associated confidence value with respect to the confidence data and determine a likely issuing state associated with the license plate.

Abstract translation: 自动确定车牌发卡状态的方法和系统。 可以通过一个或多个OCR引擎来处理由ALPR引擎获取的车牌的图像，使得OCR引擎中的每个OCR引擎被调谐到特定状态。 可以分析从OCR引擎输出的置信度数据（以及其他因素）来估计与车牌相关的发卡状态。 可以合并与发布状态有关的多个观察结果，以获得总体结论，并且相对于置信度数据分配相关的置信度值，并确定与该牌照相关联的可能发布状态。