专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明公开

US20240135139A1 Implementing Traditional Computer Vision Algorithms as Neural Networks 审中-公开
公开(公告)号：US20240135139A1
公开(公告)日：2024-04-25
申请号：US18136709
申请日：2023-04-19
申请人： Imagination Technologies Limited
发明人： Paul Brasnett , Daniel Valdez Balderas , Cagatay Dikici , Szabolcs Csefalvay , David Hough , Timothy Smith , James Imber
IPC分类号： G06N3/02 , G06F17/15 , G06F17/16 , G06N3/063 , G06N3/08
CPC分类号： G06N3/02 , G06F17/15 , G06F17/16 , G06N3/063 , G06N3/08
摘要： Methods and systems for implementing a traditional computer vision algorithm as a neural network. The method includes: receiving a definition of the traditional computer vision algorithm that identifies a sequence of one or more traditional computer vision algorithm operations; mapping each of the one or more traditional computer vision algorithm operations to a set of one or more neural network primitives that is mathematically equivalent to that traditional computer vision algorithm operation; linking the one or more network primitives mapped to each traditional computer vision algorithm operation according to the sequence to form a neural network representing the traditional computer vision algorithm; and configuring hardware logic capable of implementing a neural network to implement the neural network that represents the traditional computer vision algorithm.

2. 发明授权

US11282216B2 Image noise reduction 有权
公开(公告)号：US11282216B2
公开(公告)日：2022-03-22
申请号：US16865884
申请日：2020-05-04
申请人： Imagination Technologies Limited
发明人： Marc Vivet , Paul Brasnett
IPC分类号： G06T7/30 , G06T5/50 , G06T5/00 , G06T7/32 , G06K9/62
摘要： A reduced noise image can be formed from a set of images. One of the images of the set can be selected to be a reference image and other images of the set are transformed such that they are better aligned with the reference image. A measure of the alignment of each image with the reference image is determined. At least some of the transformed images can then be combined using weights which depend on the alignment of the transformed image with the reference image to thereby form the reduced noise image. By weighting the images according to their alignment with the reference image the effects of misalignment between the images in the combined image are reduced. Furthermore, motion correction may be applied to the reduced noise image.

3. 发明申请

US20210073614A1 METHODS AND SYSTEMS FOR CONVERTING WEIGHTS OF A DEEP NEURAL NETWORK FROM A FIRST NUMBER FORMAT TO A SECOND NUMBER FORMAT 有权
公开(公告)号：US20210073614A1
公开(公告)日：2021-03-11
申请号：US17000468
申请日：2020-08-24
申请人： Imagination Technologies Limited
发明人： Cagatay Dikici , Paul Brasnett , Muhammad Asad , Stephen Morphet
IPC分类号： G06N3/04 , G06N3/08
摘要： Methods and system for converting a plurality of weights of a filter of a Deep Neural Network (DNN) in a first number format to a second number format, the second number format having less precision than the first number format, to enable the DNN to be implemented in hardware logic. The method comprising: determining, for each of the plurality of weights, a quantisation error associated with quantising that weight to the second number format in accordance with a first quantisation method; determining a total quantisation error for the plurality of weights based on the quantisation errors for the plurality of weights; identifying a subset of the plurality of weights to be quantised to the second number format in accordance with a second quantisation method based on the total quantisation error for the plurality of weights; and generating a set of quantised weights representing the plurality of weights in the second number format, the quantised weight for each weight in the subset of the plurality of weights based on quantising that weight to the second number format in accordance with the second quantisation method and the quantised weight for each of the remaining weights of the plurality of weights based on quantising that weight to the second number format in accordance with the first quantisation method.

4. 发明申请

US20190026857A1 Upscaling Lower Resolution Image Data for Processing 审中-公开
公开(公告)号：US20190026857A1
公开(公告)日：2019-01-24
申请号：US16138302
申请日：2018-09-21
申请人： Imagination Technologies Limited
发明人： Paul Brasnett , Jonathan Diggins , Steven Fishwick , Stephen Morphet
IPC分类号： G06T1/60 , H04N19/426 , G06T3/00 , H04N19/33
摘要： In an example method and system, image data to an image processing module. Image data is read from memory into a down-scaler, which down-scales the image data to a first resolution, which is stored in a first buffer. A region of image data which the image processing module will request is predicted, and image data corresponding to at least part of the predicted region of image data is stored in a first buffer, in a second resolution, higher than the first. When a request for image data is received, it is then determined whether image data corresponding to the requested image data is in the second buffer, and if so, then image data is provided to the image processing module from the second buffer. If not, then image data from the first buffer is up-scaled, and the up-scaled image data is provided to the image processing module.

5. 发明申请

US20180150959A1 IMAGE NOISE REDUCTION 审中-公开
公开(公告)号：US20180150959A1
公开(公告)日：2018-05-31
申请号：US15877552
申请日：2018-01-23
申请人： Imagination Technologies Limited
发明人： Marc Vivet , Paul Brasnett
IPC分类号： G06T7/30 , G06T5/00 , G06T5/50 , G06T7/32 , G06K9/62
摘要： A reduced noise image can be formed from a set of images. One of the images of the set can be selected to be a reference image and other images of the set are transformed such that they are better aligned with the reference image. A measure of the alignment of each image with the reference image is determined. At least some of the transformed images can then be combined using weights which depend on the alignment of the transformed image with the reference image to thereby form the reduced noise image. By weighting the images according to their alignment with the reference image the effects of misalignment between the images in the combined image are reduced. Furthermore, motion correction may be applied to the reduced noise image.

6. 发明授权

US09633249B2 Face detection in an image data stream using skin colour patch constructs 有权
公开(公告)号：US09633249B2
公开(公告)日：2017-04-25
申请号：US14521957
申请日：2014-10-23
申请人： Imagination Technologies Limited
发明人： Szabolcs Cséfalvay , Paul Brasnett
IPC分类号： G06K9/00 , G06T7/00
CPC分类号： G06K9/00234 , G06K9/00228 , G06T7/11 , G06T7/75 , G06T7/90 , G06T2207/10016 , G06T2207/10024 , G06T2207/30201
摘要： A data processing system for performing face detection on a stream of frames of image data, the data processing system comprising: a skin patch identifier configured to identify one or more patches of skin color in a first frame and characterize each patch in the first frame using a respective patch construct of a predefined shape; a first search tile generator configured to generate one or more first search tiles from the one or more patch constructs; and a face detector configured to detect faces in the stream by performing face detection in one or more frames of the stream within the first search tiles.

7. 发明授权

US09349037B2 Skin colour probability map 有权
标题翻译：皮肤颜色概率图
公开(公告)号：US09349037B2
公开(公告)日：2016-05-24
申请号：US14522009
申请日：2014-10-23
申请人： Imagination Technologies Limited
发明人： Szabolcs Cséfalvay , Paul Brasnett
IPC分类号： G06K9/00 , G06T7/00
CPC分类号： G06K9/00234 , G06K9/00228 , G06T7/11 , G06T7/75 , G06T7/90 , G06T2207/10016 , G06T2207/10024 , G06T2207/30201
摘要： A data processing system for performing face detection on a stream of frames of image data, the data processing system comprising: a face detector configured to detect a first face candidate in a first frame by performing face detection within first search tiles defined for the first frame; a color measurement unit configured to calculate a set of color parameters including an average color of the first face candidate expressed according to a predefined color space; a transformation unit configured to: transform a second frame into the predefined color space, one of the axes of the color space being substantially oriented in the direction of maximum variation according to a predetermined distribution of skin color; and form a skin color probability map for the second frame by calculating the probability that a given color is a skin color from a measure of the color space distance of that color from the calculated average color; and a search tile generator configured to generate second search tiles based on the skin color probability map for use by the face detector, the second search tiles defining areas of the second frame within which the face detector is to perform face detection so as to detect one or more second face candidates in the second frame.
摘要翻译：一种用于对图像数据帧进行面部检测的数据处理系统，所述数据处理系统包括：面部检测器，被配置为通过在为所述第一帧定义的第一搜索瓦片内进行面部检测来检测第一帧中的第一面部候选 ; 颜色测量单元，被配置为计算包括根据预定义的颜色空间表达的第一面部候选者的平均颜色的一组颜色参数; 变换单元，被配置为：将第二帧变换成所述预定颜色空间，所述颜色空间的轴中的一个基本上根据皮肤颜色的预定分布沿最大变化的方向定向; 并且通过从所计算的平均颜色的颜色空间距离的测量中计算给定颜色是肤色的概率来形成第二帧的肤色概率图; 以及搜索瓦片发生器，被配置为基于面部检测器使用的皮肤颜色概率图来生成第二搜索瓦片，所述第二搜索瓦片定义面部检测器将进行面部检测的第二框架的区域，以检测一个或更多第二面部候选者。

8. 发明授权

US11587199B2 Upscaling lower resolution image data for processing 有权
公开(公告)号：US11587199B2
公开(公告)日：2023-02-21
申请号：US16138302
申请日：2018-09-21
申请人： Imagination Technologies Limited
发明人： Paul Brasnett , Jonathan Diggins , Steven Fishwick , Stephen Morphet
IPC分类号： G06T1/60 , G06T3/00 , H04N19/33 , H04N19/426
摘要： In an example method and system, image data to an image processing module. Image data is read from memory into a down-scaler, which down-scales the image data to a first resolution, which is stored in a first buffer. A region of image data which the image processing module will request is predicted, and image data corresponding to at least part of the predicted region of image data is stored in a first buffer, in a second resolution, higher than the first. When a request for image data is received, it is then determined whether image data corresponding to the requested image data is in the second buffer, and if so, then image data is provided to the image processing module from the second buffer. If not, then image data from the first buffer is up-scaled, and the up-scaled image data is provided to the image processing module.

9. 发明申请

US20220067497A1 METHODS AND SYSTEMS FOR CONVERTING WEIGHTS OF A DEEP NEURAL NETWORK FROM A FIRST NUMBER FORMAT TO A SECOND NUMBER FORMAT 有权
公开(公告)号：US20220067497A1
公开(公告)日：2022-03-03
申请号：US17524606
申请日：2021-11-11
申请人： Imagination Technologies Limited
发明人： Cagatay Dikici , Paul Brasnett , Muhammad Asad , Stephen Morphet
IPC分类号： G06N3/063 , G06N3/04 , H03M7/24
摘要： Methods and system for converting a plurality of weights of a filter of a Deep Neural Network (DNN) in a first number format to a second number format, the second number format having less precision than the first number format, to enable the DNN to be implemented in hardware logic. The method comprising: determining, for each of the plurality of weights, a quantisation error associated with quantising that weight to the second number format in accordance with a first quantisation method; determining a total quantisation error for the plurality of weights based on the quantisation errors for the plurality of weights; identifying a subset of the plurality of weights to be quantised to the second number format in accordance with a second quantisation method based on the total quantisation error for the plurality of weights; and generating a set of quantised weights representing the plurality of weights in the second number format, the quantised weight for each weight in the subset of the plurality of weights based on quantising that weight to the second number format in accordance with the second quantisation method and the quantised weight for each of the remaining weights of the plurality of weights based on quantising that weight to the second number format in accordance with the first quantisation method.

10. 发明授权

US11188817B2 Methods and systems for converting weights of a deep neural network from a first number format to a second number format 有权
公开(公告)号：US11188817B2
公开(公告)日：2021-11-30
申请号：US17000468
申请日：2020-08-24
申请人： Imagination Technologies Limited
发明人： Cagatay Dikici , Paul Brasnett , Muhammad Asad , Stephen Morphet
IPC分类号： G06N3/063 , G06N3/04 , H03M7/24 , H03M7/30
摘要： Methods and system for converting a plurality of weights of a filter of a Deep Neural Network (DNN) in a first number format to a second number format, the second number format having less precision than the first number format, to enable the DNN to be implemented in hardware logic. The method comprising: determining, for each of the plurality of weights, a quantisation error associated with quantising that weight to the second number format in accordance with a first quantisation method; determining a total quantisation error for the plurality of weights based on the quantisation errors for the plurality of weights; identifying a subset of the plurality of weights to be quantised to the second number format in accordance with a second quantisation method based on the total quantisation error for the plurality of weights; and generating a set of quantised weights representing the plurality of weights in the second number format, the quantised weight for each weight in the subset of the plurality of weights based on quantising that weight to the second number format in accordance with the second quantisation method and the quantised weight for each of the remaining weights of the plurality of weights based on quantising that weight to the second number format in accordance with the first quantisation method.

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式