专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

21. 发明公开

US20230377317A1 SYSTEMS AND METHODS FOR INTELLIGENT SELECTION OF DATA FOR BUILDING A MACHINE LEARNING MODEL 审中-公开
公开(公告)号：US20230377317A1
公开(公告)日：2023-11-23
申请号：US18364569
申请日：2023-08-03
申请人： Argo AI, LLC
发明人： Jelena Frtunikj , Daniel Alfonsetti
IPC分类号： G06V10/774 , G06N20/00 , G06V10/20 , G06V20/56 , G06F18/214 , G06F18/2115 , G06V10/772 , G06V10/82 , G06V20/58
CPC分类号： G06V10/774 , G06N20/00 , G06V10/255 , G06V20/56 , G06F18/2148 , G06F18/2115 , G06V10/772 , G06V10/82 , G06V20/58 , G06V20/588 , G06V20/64
摘要： Systems and methods for selecting data for training a machine learning model using active learning are disclosed. The methods include receiving a plurality of unlabeled sensor data logs corresponding to surroundings of an autonomous vehicle and identifying one or more trends associated with a training dataset comprising a plurality of labeled data logs. The methods also include selecting a subset of the plurality of unlabeled sensor data logs that have an importance score greater than a threshold, the importance score being determined based on the one or more trends. The subset of the plurality of unlabeled sensor data logs is used for updating the machine learning model to generate an updated model.

22. 发明授权

US11823428B2 Image processing apparatus and control method therefor, image capturing apparatus, and storage medium 有权
公开(公告)号：US11823428B2
公开(公告)日：2023-11-21
申请号：US17173810
申请日：2021-02-11
申请人： CANON KABUSHIKI KAISHA
发明人： Ryota Kashu
IPC分类号： G06V10/25 , G06T7/246 , G06V10/772 , G06F18/28 , G06F18/24 , G06F18/25 , G06F18/40 , G06F18/20 , H04N23/61 , H04N23/67
CPC分类号： G06V10/25 , G06F18/24 , G06F18/25 , G06F18/28 , G06F18/285 , G06F18/40 , G06T7/248 , G06V10/772 , H04N23/61 , H04N23/675
摘要： An apparatus comprises a storage device configured to store a plurality of dictionary data for respectively detecting a plurality of different objects from an image; and at least one processor configured to function as: a detection unit configured to use partial dictionary data of the plurality of dictionary data to detect an object corresponding to the partial dictionary data, with respect to each frame of an image of a plurality of frames obtained by an image capturing device; and a switching unit configured to switch the dictionary data to be used by the detection unit in the plurality of frames, according to a result of the object detection by the detection unit.

23. 发明公开

US20230368501A1 FEW-SHOT TRAINING OF A NEURAL NETWORK 审中-公开
公开(公告)号：US20230368501A1
公开(公告)日：2023-11-16
申请号：US18114177
申请日：2023-02-24
申请人： NVIDIA Corporation
发明人： Seonwook Park , Shalini De Mello , Pavlo Molchanov , Umar Iqbal , Jan Kautz
IPC分类号： G06V10/772 , G06F7/57 , G06F17/18 , G06N3/088 , G06N3/045 , G06N3/047 , G06V10/774 , G06V10/82
CPC分类号： G06V10/772 , G06F7/57 , G06F17/18 , G06N3/088 , G06N3/045 , G06N3/047 , G06V10/774 , G06V10/82
摘要： A neural network is trained to identify one or more features of an image. The neural network is trained using a small number of original images, from which a plurality of additional images are derived. The additional images generated by rotating and decoding embeddings of the image in a latent space generated by an autoencoder. The images generated by the rotation and decoding exhibit changes to a feature that is in proportion to the amount of rotation.

24. 发明授权

US11790066B2 Systems and methods for private authentication with helper networks 有权
公开(公告)号：US11790066B2
公开(公告)日：2023-10-17
申请号：US17473360
申请日：2021-09-13
申请人： Private Identity LLC
发明人： Scott Edward Streit
IPC分类号： H04L29/06 , G06N3/04 , G06K9/62 , G06F21/32 , H04L9/40 , G06F18/214 , G06V10/772 , G06V10/774 , G06V10/82 , G06V10/98 , G06V40/12 , G06V40/16 , G06V40/40
CPC分类号： G06F21/32 , G06F18/214 , G06N3/04 , G06V10/772 , G06V10/774 , G06V10/82 , G06V10/993 , G06V40/12 , G06V40/16 , G06V40/40 , H04L63/0236 , H04L63/1416 , H04L63/1466
摘要： Helper neural network can play a role in augmenting authentication services that are based on neural network architectures. For example, helper networks are configured to operate as a gateway on identification information used to identify users, enroll users, and/or construct authentication models (e.g., embedding and/or prediction networks). Assuming, that both good and bad identification information samples are taken as part of identification information capture, the helper networks operate to filter out bad identification information prior to training, which prevents, for example, identification information that is valid but poorly captured from impacting identification, training, and/or prediction using various neural networks. Additionally, helper networks can also identify and prevent presentation attacks or submission of spoofed identification information as part of processing and/or validation.

25. 发明公开

US20230274532A1 Learning Data Augmentation Strategies for Object Detection 审中-公开
公开(公告)号：US20230274532A1
公开(公告)日：2023-08-31
申请号：US18313772
申请日：2023-05-08
申请人： Google LLC
发明人： Jon Shlens , Ekin Dogus Cubuk , Quoc Le , Tsung-Yi Lin , Barret Zoph , Golnaz Ghiasi
IPC分类号： G06V10/772 , G06T3/60 , G06T3/20 , G06T11/00 , G06F18/24 , G06F18/21
CPC分类号： G06V10/772 , G06F18/217 , G06F18/24 , G06T3/20 , G06T3/60 , G06T11/001
摘要： Example aspects of the present disclosure are directed to systems and methods for learning data augmentation strategies for improved object detection model performance. In particular, example aspects of the present disclosure are directed to iterative reinforcement learning approaches in which, at each of a plurality of iterations, a controller model selects a series of one or more augmentation operations to be applied to training images to generate augmented images. For example, the controller model can select the augmentation operations from a defined search space of available operations which can, for example, include operations that augment the training image without modification of the locations of a target object and corresponding bounding shape within the image and/or operations that do modify the locations of the target object and bounding shape within the training image.

26. 发明公开

US20230252614A1 IMAGE HARMONIZATION FOR DEEP LEARNING MODEL OPTIMIZATION 审中-公开
公开(公告)号：US20230252614A1
公开(公告)日：2023-08-10
申请号：US18304947
申请日：2023-04-21
申请人： GE Precision Healthcare LLC
发明人： Tao Tan , Pál Tegzes , Levente Imre Török , Lehel Ferenczi , Gopal B. Avinash , László Ruskó , Gireesha Chinthamani Rao , Khaled Younis , Soumya Ghose
IPC分类号： G06T5/50 , G06V10/772 , G06V10/774 , G06V10/762 , G06V10/74 , G06V10/776 , G06T7/00 , G06V10/82 , G06F18/21 , G06F18/22 , G06F18/23 , G06F18/28 , G06F18/214
CPC分类号： G06T5/50 , G06V10/772 , G06V10/774 , G06V10/762 , G06V10/761 , G06V10/776 , G06T7/00 , G06V10/82 , G06F18/217 , G06F18/22 , G06F18/23 , G06F18/28 , G06F18/214
摘要： Techniques are described for optimizing deep learning model performance using image harmonization as a pre-processing step. According to an embodiment, a method comprises decomposing, by a system operatively coupled to a processor, an input image into sub-images. The method further comprises harmonizing the sub-images with corresponding reference sub-images of at least one reference image based on two or more different statistical values respectively calculated for the sub-images and the corresponding reference-sub images, resulting in transformation of the sub-images into modified sub-images images. In some implementations, the modified sub-images can be combined into a harmonized image having a more similar appearance to the at least one reference image relative to the input image. In other implementations, harmonized images and/or modified sub-images generated using these techniques can be used as ground-truth training samples for training one or more deep learning model to transform input images with appearance variations into harmonized images.

27. 发明授权

US11669960B2 Learning system, method, and program 有权
公开(公告)号：US11669960B2
公开(公告)日：2023-06-06
申请号：US16927956
申请日：2020-07-13
申请人： FUJIFILM Corporation
发明人： Akimichi Ichinose , Keigo Nakamura
IPC分类号： G06T5/50 , G06F18/22 , G06V10/772 , G06T7/00 , G06F18/214 , G06V10/774 , G06V10/776 , G06V10/82
CPC分类号： G06T7/0012 , G06F18/214 , G06F18/22 , G06T5/50 , G06V10/772 , G06V10/774 , G06V10/776 , G06V10/82 , G06T2207/20081
摘要： Provided are a learning system, a method, and a program that can provide medical institutions with motivation for creating learning data related to medical images in a case of using machine learning to support diagnosis using medical images. The learning system includes a reception unit that receives an input of first learning data from a user, a calculation unit that calculates, a contribution degree of the first learning data to learning of a discriminator for each user on the basis of at least one of a comparison result between the first learning data and second learning data used for creating the discriminator or a comparison result between an output obtained by inputting the first learning data to the discriminator and correct answer data corresponding to the first learning data, and a service setting unit that sets a service for the user on the basis of the contribution degree calculated for each user.

28. 发明授权

US12062225B2 Deep neural network visualisation 有权
公开(公告)号：US12062225B2
公开(公告)日：2024-08-13
申请号：US17615946
申请日：2020-05-25
申请人： KONINKLIJKE PHILIPS N.V.
发明人： Bart Jacob Bakker , Dimitrios Mavroeidis , Stojan Trajanovski
IPC分类号： G06V10/764 , G06V10/772 , G06V10/774 , G06V10/82
CPC分类号： G06V10/764 , G06V10/772 , G06V10/774 , G06V10/82
摘要： Aspects and embodiments relate to a method of providing a representation of a feature identified by a deep neural network as being relevant to an outcome, a computer program product and apparatus configured to perform that method. The method comprises: providing the deep neural network with a training library comprising: a plurality of samples associated with the outcome; using the deep neural network to recognise a feature in the plurality of samples associated with the outcome; creating a feature recognition library from an input library by identifying one or more elements in each of a plurality of samples in the input library which trigger recognition of the feature by the deep neural network; using the feature recognition library to synthesise a plurality of one or more elements of a sample which have characteristics which trigger recognition of the feature by the deep neural network; and using the synthesised plurality of one or more elements to provide a representation of the feature identified by the deep neural network in the plurality of samples associated with the outcome. Accordingly, rather than visualising a single instance of one or more elements in a sample which trigger a feature associated with an outcome, it is possible to visualise a range of samples including elements which would trigger a feature associated with an outcome, thus enabling a more comprehensive view of operation of a deep neural network in relation to a particular feature.

29. 发明授权

US12051178B2 Image harmonization for deep learning model optimization 有权
公开(公告)号：US12051178B2
公开(公告)日：2024-07-30
申请号：US18304947
申请日：2023-04-21
申请人： GE Precision Healthcare LLC
发明人： Tao Tan , Pál Tegzes , Levente Imre Török , Lehel Ferenczi , Gopal B. Avinash , László Ruskó , Gireesha Chinthamani Rao , Khaled Younis , Soumya Ghose
IPC分类号： G06T5/50 , G06F18/21 , G06F18/214 , G06F18/22 , G06F18/23 , G06F18/28 , G06T7/00 , G06V10/74 , G06V10/762 , G06V10/772 , G06V10/774 , G06V10/776 , G06V10/82
CPC分类号： G06T5/50 , G06F18/214 , G06F18/217 , G06F18/22 , G06F18/23 , G06F18/28 , G06T7/00 , G06V10/761 , G06V10/762 , G06V10/772 , G06V10/774 , G06V10/776 , G06V10/82
摘要： Techniques are described for optimizing deep learning model performance using image harmonization as a pre-processing step. According to an embodiment, a method comprises decomposing, by a system operatively coupled to a processor, an input image into sub-images. The method further comprises harmonizing the sub-images with corresponding reference sub-images of at least one reference image based on two or more different statistical values respectively calculated for the sub-images and the corresponding reference-sub images, resulting in transformation of the sub-images into modified sub-images images. In some implementations, the modified sub-images can be combined into a harmonized image having a more similar appearance to the at least one reference image relative to the input image. In other implementations, harmonized images and/or modified sub-images generated using these techniques can be used as ground-truth training samples for training one or more deep learning model to transform input images with appearance variations into harmonized images.

30. 发明公开

US20240233177A9 SYSTEM AND METHOD FOR EFFICIENTLY SCORING PROBES IN AN IMAGE WITH A VISION SYSTEM 审中-公开
公开(公告)号：US20240233177A9
公开(公告)日：2024-07-11
申请号：US18208660
申请日：2023-06-12
申请人： Cognex Corporation
发明人： Nathaniel R. Bogan
IPC分类号： G06T7/73 , G06F18/28 , G06T7/00 , G06T7/33 , G06V10/44 , G06V10/46 , G06V10/75 , G06V10/772
CPC分类号： G06T7/75 , G06F18/28 , G06T7/001 , G06T7/344 , G06V10/44 , G06V10/46 , G06V10/752 , G06V10/772 , G06T2207/20081 , G06T2207/30164 , G06V2201/06
摘要： A system and method for scoring trained probes for use in analyzing one or more candidate poses of a runtime image is provided. A set of probes with location and gradient direction based on a trained model are applied to one or more candidate poses based upon a runtime image. The applied probes each respectively include a discrete set of position offsets with respect to the gradient direction thereof. A match score is computed for each of the probes, which includes estimating a best match position for each of the probes respectively relative to one of the offsets thereof, and generating a set of individual probe scores for each of the probes, respectively at the estimated best match position.

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