公开(公告)号：US11042775B1

公开(公告)日：2021-06-22

申请号：US15187533

申请日：2016-06-20

申请人： BRAIN Corporation

发明人： Micah Richert ,  Filip Piekniewski

IPC分类号： G06K9/62 ,  G06N3/04

摘要： A data processing apparatus may utilize an artificial neuron network configured to reduce dimensionality of input data using a sparse transformation configured using receptive field structure of network units. Output of the network may be analyzed for temporally persistency that is characterized by similarity matrix. Elements of the matrix may be incremented when present activity unit activity at a preceding frame. The similarity matrix may be partitioned based on a distance measure for a given element of the matrix and its closest neighbors. Stability of learning of temporally proximal patterns may be greatly improved as the similarity matrix is learned independently of the partitioning operation. Partitioning of the similarity matrix using the methodology of the disclosure may be performed online, e.g., contemporaneously with the encoding and/or similarity matrix construction, thereby enabling learning of new features in the input data.

公开(公告)号：US20190005659A1

公开(公告)日：2019-01-03

申请号：US16104622

申请日：2018-08-17

申请人： BRAIN CORPORATION

发明人： Filip Piekniewski ,  Micah Richert ,  Dimitry Fisher

IPC分类号： G06T7/292

摘要： Apparatus and methods for detecting and utilizing saliency in digital images. In one implementation, salient objects may be detected based on analysis of pixel characteristics. Least frequently occurring pixel values may be deemed as salient. Pixel values in an image may be compared to a reference. Color distance may be determined based on a difference between reference color and pixel color. Individual image channels may be scaled when determining saliency in a multi-channel image. Areas of high saliency may be analyzed to determine object position, shape, and/or color. Multiple saliency maps may be additively or multiplicative combined in order to improve detection performance (e.g., reduce number of false positives). Methodologies described herein may enable robust tracking of objects utilizing fewer determination resources. Efficient implementation of the methods described below may allow them to be used for example on board a robot (or autonomous vehicle) or a mobile determining platform.

公开(公告)号：US09373038B2

公开(公告)日：2016-06-21

申请号：US14191383

申请日：2014-02-26

申请人： BRAIN CORPORATION

发明人： Micah Richert ,  Filip Piekniewski

IPC分类号： G06K9/00 ,  G06N3/08 ,  G06N3/04 ,  G06K9/46 ,  G06T7/20

CPC分类号： G06K9/00711 ,  G06K9/4671 ,  G06N3/049 ,  G06N3/08 ,  G06T7/246 ,  G06T2207/10016

摘要： A data processing apparatus may utilize an artificial neuron network configured to reduce dimensionality of input data using a sparse transformation configured using receptive field structure of network units. Output of the network may be analyzed for temporally persistency that is characterized by similarity matrix. Elements of the matrix may be incremented when present activity unit activity at a preceding frame. The similarity matrix may be partitioned based on a distance measure for a given element of the matrix and its closest neighbors. Stability of learning of temporally proximal patterns may be greatly improved as the similarity matrix is learned independently of the partitioning operation. Partitioning of the similarity matrix using the methodology of the disclosure may be performed online, e.g., contemporaneously with the encoding and/or similarity matrix construction, thereby enabling learning of new features in the input data.

摘要翻译： 数据处理装置可以使用被配置为使用使用网络单元的接收场结构配置的稀疏变换来减少输入数据的维度的人造神经元网络。 可以通过相似矩阵来表征网络的时间持续性的输出。 当在前一帧处存在活动单元活动时，矩阵的元素可以增加。 可以基于矩阵的给定元素及其最近邻居的距离度量来划分相似性矩阵。 随着独立于划分操作学习相似矩阵，可以大大提高时间近似图案的学习的稳定性。 使用本公开的方法对相似性矩阵进行分割可以在线执行，例如与编码和/或相似矩阵构造同时进行，从而使得能够学习输入数据中的新特征。

公开(公告)号：US20160086052A1

公开(公告)日：2016-03-24

申请号：US14637191

申请日：2015-03-03

申请人： Brain Corporation

发明人： Filip Piekniewski ,  Micah Richert ,  Dimitry Fisher

IPC分类号： G06K9/46 ,  G06T7/00 ,  G06T7/20 ,  G06K9/32

CPC分类号： G06T7/292 ,  G06K9/3241 ,  G06K9/4623 ,  G06K9/4642 ,  G06K9/4652 ,  G06K9/4671 ,  G06K9/48 ,  G06T7/20 ,  G06T7/246 ,  G06T7/90 ,  G06T2207/10024 ,  G06T2207/20164 ,  G06T2207/30241 ,  G06T2215/16 ,  H04N9/7973

摘要： Apparatus and methods for detecting and utilizing saliency in digital images. In one implementation, salient objects may be detected based on analysis of pixel characteristics. Least frequently occurring pixel values may be deemed as salient. Pixel values in an image may be compared to a reference. Color distance may be determined based on a difference between reference color and pixel color. Individual image channels may be scaled when determining saliency in a multi-channel image. Areas of high saliency may be analyzed to determine object position, shape, and/or color. Multiple saliency maps may be additively or multiplicative combined in order to improve detection performance (e.g., reduce number of false positives). Methodologies described herein may enable robust tracking of objects utilizing fewer determination resources. Efficient implementation of the methods described below may allow them to be used for example on board a robot (or autonomous vehicle) or a mobile determining platform.

摘要翻译： 用于检测和利用数字图像显着性的装置和方法。 在一个实现中，可以基于像素特性的分析来检测显着对象。 最常出现的像素值可能被认为是显着的。 图像中的像素值可以与参考进行比较。 可以基于参考颜色和像素颜色之间的差来确定颜色距离。 当确定多通道图像中的显着性时，可以缩放各个图像通道。 可以分析高显着性的区域以确定对象位置，形状和/或颜色。 为了提高检测性能（例如，减少误报数量），多重显着图可以被相加或乘法组合。 本文描述的方法可以利用较少的确定资源来实现对物体的鲁棒跟踪。 下面描述的方法的有效实现可以允许它们例如在机器人（或自主车辆）或移动确定平台上使用。

公开(公告)号：US20160086050A1

公开(公告)日：2016-03-24

申请号：US14637138

申请日：2015-03-03

申请人： Brain Corporation

发明人： Filip Piekniewski ,  Micah Richert ,  Dimitry Fisher

IPC分类号： G06K9/46 ,  G06T7/40 ,  G06T3/40 ,  G06T5/40 ,  G06K9/62 ,  G06T3/00

CPC分类号： G06T7/292 ,  G06K9/3241 ,  G06K9/4623 ,  G06K9/4642 ,  G06K9/4652 ,  G06K9/4671 ,  G06K9/48 ,  G06T7/20 ,  G06T7/246 ,  G06T7/90 ,  G06T2207/10024 ,  G06T2207/20164 ,  G06T2207/30241 ,  G06T2215/16 ,  H04N9/7973

摘要： Apparatus and methods for detecting and utilizing saliency in digital images. In one implementation, salient objects may be detected based on analysis of pixel characteristics. Least frequently occurring pixel values may be deemed as salient. Pixel values in an image may be compared to a reference. Color distance may be determined based on a difference between reference color and pixel color. Individual image channels may be scaled when determining saliency in a multi-channel image. Areas of high saliency may be analyzed to determine object position, shape, and/or color. Multiple saliency maps may be additively or multiplicative combined in order to improve detection performance (e.g., reduce number of false positives). Methodologies described herein may enable robust tracking of objects utilizing fewer determination resources. Efficient implementation of the methods described below may allow them to be used for example on board a robot (or autonomous vehicle) or a mobile determining platform.

摘要翻译： 用于检测和利用数字图像显着性的装置和方法。 在一个实现中，可以基于像素特性的分析来检测显着对象。 最常出现的像素值可能被认为是显着的。 图像中的像素值可以与参考进行比较。 可以基于参考颜色和像素颜色之间的差来确定颜色距离。 当确定多通道图像中的显着性时，可以缩放各个图像通道。 可以分析高显着性的区域以确定对象位置，形状和/或颜色。 为了提高检测性能（例如，减少误报数量），多重显着图可以被相加或乘法组合。 本文描述的方法可以利用较少的确定资源来实现对物体的鲁棒跟踪。 下面描述的方法的有效实现可以允许它们例如在机器人（或自主车辆）或移动确定平台上使用。

公开(公告)号：US20150242690A1

公开(公告)日：2015-08-27

申请号：US14191383

申请日：2014-02-26

申请人： BRAIN CORPORATION

发明人： Micah Richert ,  Filip Piekniewski

IPC分类号： G06K9/00 ,  G06K9/46 ,  G06T7/20

CPC分类号： G06K9/00711 ,  G06K9/4671 ,  G06N3/049 ,  G06N3/08 ,  G06T7/246 ,  G06T2207/10016

摘要： A data processing apparatus may utilize an artificial neuron network configured to reduce dimensionality of input data using a sparse transformation configured using receptive field structure of network units. Output of the network may be analyzed for temporally persistency that is characterized by similarity matrix. Elements of the matrix may be incremented when present activity unit activity at a preceding frame. The similarity matrix may be partitioned based on a distance measure for a given element of the matrix and its closest neighbors. Stability of learning of temporally proximal patterns may be greatly improved as the similarity matrix is learned independently of the partitioning operation. Partitioning of the similarity matrix using the methodology of the disclosure may be performed online, e.g., contemporaneously with the encoding and/or similarity matrix construction, thereby enabling learning of new features in the input data.

摘要翻译： 数据处理装置可以使用被配置为使用使用网络单元的接收场结构配置的稀疏变换来减少输入数据的维度的人造神经元网络。 可以通过相似矩阵来表征网络的时间持续性的输出。 当在前一帧处存在活动单元活动时，矩阵的元素可以增加。 可以基于矩阵的给定元素及其最近邻居的距离度量来划分相似性矩阵。 随着独立于划分操作学习相似矩阵，可以大大提高时间近似图案的学习的稳定性。 使用本公开的方法对相似性矩阵进行分割可以在线执行，例如与编码和/或相似矩阵构造同时进行，从而使得能够学习输入数据中的新特征。

公开(公告)号：US10728436B2

公开(公告)日：2020-07-28

申请号：US15845891

申请日：2017-12-18

申请人： Brain Corporation

发明人： Filip Piekniewski ,  Vadim Polonichko ,  Eugene Izhikevich

IPC分类号： H04N5/225 ,  B60W30/09 ,  B60W30/16 ,  G06K9/00 ,  G02B5/20 ,  G02B13/00 ,  G02B27/00 ,  B60W30/14 ,  G01C3/08 ,  G02B7/38 ,  G06K9/46 ,  G06K9/52 ,  H01L31/0216

摘要： An optical object detection apparatus and associated methods. The apparatus may comprise a lens (e.g., fixed-focal length wide aperture lens) and an image sensor. The fixed focal length of the lens may correspond to a depth of field area in front of the lens. When an object enters the depth of field area (e.g., sue to a relative motion between the object and the lens) the object representation on the image sensor plane may be in-focus. Objects outside the depth of field area may be out of focus. In-focus representations of objects may be characterized by a greater contrast parameter compared to out of focus representations. One or more images provided by the detection apparatus may be analyzed in order to determine useful information (e.g., an image contrast parameter) of a given image. Based on the image contrast meeting one or more criteria, a detection indication may be produced.

公开(公告)号：US10464213B2

公开(公告)日：2019-11-05

申请号：US15997397

申请日：2018-06-04

申请人： BRAIN CORPORATION

发明人： Dimitry Fisher ,  Cody Griffin ,  Micah Richert ,  Filip Piekniewski ,  Eugene Izhikevich ,  Jayram Moorkanikara Nageswaran ,  John Black

IPC分类号： G06K9/00 ,  B25J9/16 ,  G06K9/46 ,  G06T7/11 ,  H04N5/33 ,  G05D1/00 ,  G08B21/20

摘要： Systems and methods for automatic detection of spills are disclosed. In some exemplary implementations, a robot can have a spill detector comprising at least one optical imaging device configured to capture at least one image of a scene containing a spill while the robot moves between locations. The robot can process the at least one image by segmentation. Once the spill has been identified, the robot can then generate an alert indicative at least in part of a recognition of the spill.

公开(公告)号：US09552546B1

公开(公告)日：2017-01-24

申请号：US13954575

申请日：2013-07-30

申请人： Brain Corporation

发明人： Filip Piekniewski

IPC分类号： G06F15/18 ,  G06N3/08

CPC分类号： G06N3/08 ,  G06N3/049 ,  G06N3/063

摘要： Apparatus and methods for plasticity in spiking neuron networks. In various implementations, the efficacy of one or more connections of the network may be adjusted based on a plasticity rule during network operation. The rule may comprise a connection depression portion and/or a potentiation portion. Statistical parameters of the adjusted efficacy of a population of connections may be determined. The statistical parameter(s) may be utilized to adapt the plasticity rule during network operation in order to obtain efficacy characterized by target statistics. Based on the statistical parameter exceeding a target value, the depression magnitude of the plasticity rule may be reduced. Based on a statistical parameter being below the target value, the depression magnitude of the plasticity rule may be increased. The use of adaptive modification of the plasticity rule may improve network convergence while alleviating a need for manual tuning of efficacy during network operation.

摘要翻译： 尖峰神经元网络中可塑性的装置和方法。 在各种实现中，可以基于网络操作期间的可塑性规则来调整网络的一个或多个连接的功效。 规则可以包括连接凹陷部分和/或增强部分。 可以确定连接群体的调整功效的统计参数。 可以利用统计参数来适应网络运行期间的可塑性规则，以获得以目标统计为特征的功能。 基于超过目标值的统计参数，可塑性规则的抑制量可以减小。 基于低于目标值的统计参数，可塑性规则的抑制量可以增加。 使用可塑性规则的自适应修改可以提高网络收敛性，同时减轻在网络操作期间手动调整功效的需要。

公开(公告)号：US09436909B2

公开(公告)日：2016-09-06

申请号：US13922143

申请日：2013-06-19

申请人： Brain Corporation

发明人： Filip Piekniewski ,  Vadim Polonichko ,  Eugene Izhikevich

IPC分类号： G06N5/00 ,  G06F1/00 ,  G06N3/08 ,  G06N3/04

CPC分类号： G06N3/08 ,  G06N3/049

摘要： Apparatus and methods for processing inputs by one or more neurons of a network. The neuron(s) may generate spikes based on receipt of multiple inputs. Latency of spike generation may be determined based on an input magnitude. Inputs may be scaled using for example a non-linear concave transform. Scaling may increase neuron sensitivity to lower magnitude inputs, thereby improving latency encoding of small amplitude inputs. The transformation function may be configured compatible with existing non-scaling neuron processes and used as a plug-in to existing neuron models. Use of input scaling may allow for an improved network operation and reduce task simulation time.

摘要翻译： 用于由网络的一个或多个神经元处理输入的装置和方法。 基于多个输入的接收，神经元可以产生尖峰。 可以基于输入幅度来确定尖峰生成的延迟。 可以使用例如非线性凹变换来缩放输入。 缩放可以将神经元灵敏度增加到较低幅度的输入，从而改善小振幅输入的延迟编码。 转换函数可以被配置为与现有的非缩放神经元过程兼容，并且用作现有神经元模型的插件。 使用输入缩放可以允许改进的网络操作并减少任务模拟时间。