-
公开(公告)号:US11003959B1
公开(公告)日:2021-05-11
申请号:US16440384
申请日:2019-06-13
Applicant: AMAZON TECHNOLOGIES, INC.
Inventor: Ilya Levner , Konstantinos Boulis , Gurbinder Gill , Canku Calargun , Prajwal Yadapadithaya , Venkata Krishnan Ramamoorthy , Zhaoqing Ma
Abstract: Categorizing images may include training a first neural network to cluster a plurality of images to obtain a first image embedding space, wherein a vector representation is determined for each of the plurality of images based on the training, determining a vector norm value corresponding to each of the plurality of images based on the vector representation for each of the plurality of images, and identifying a first subset of the images for which a corresponding vector norm value satisfies a predetermined vector norm quality threshold. Then, a second neural network may be trained using the first subset of images to obtain a second image embedding space, and the second image embedding space may be used to categorize additional images.
-
公开(公告)号:US11755758B1
公开(公告)日:2023-09-12
申请号:US15798050
申请日:2017-10-30
Applicant: Amazon Technologies, Inc.
Inventor: Scott Southwood , Canku Alp Calargun , Brett Lounsbury , Prajwal Yadapadithaya , Ankit Garg , Chih-Jen Huang , Vivek Chaganti
IPC: G06F21/62 , G06V30/413 , G06V30/19 , G06V30/414 , G06V30/418
CPC classification number: G06F21/6218 , G06V30/19073 , G06V30/413 , G06V30/414 , G06V30/418
Abstract: Various embodiments provide processes for identification of embedded unauthorized data within document data. Such processes may include evaluating document data properties for indications of closing headers, comparing document data sizes against reference sizes, and comparing document data sizes against other document data sizes. Certain embodiments evaluating image data may evaluate image properties, such as histograms or gradient images, to determine random or otherwise unauthorized data. Upon detection of unauthorized data, the document data may be flagged for later processing, such as removal from a storage server.
-
公开(公告)号:US11354936B1
公开(公告)日:2022-06-07
申请号:US16929387
申请日:2020-07-15
Applicant: Amazon Technologies, Inc.
Inventor: Dharmil Satishbhai Chandarana , Ilya Levner , Zhaoqing Ma , Prajwal Yadapadithaya , Riley James Williams , Canku Alp Calargun , Prama Anand
Abstract: Techniques for improved image classification are provided. Face embeddings are generated for each face depicted in a collection of images, and the face embeddings are clustered based on the individual whose face is depicted. Based on these clusters, each embedding is assigned a label reflecting the cluster assignments. Some or all of the face embeddings are then used to train a classifier model to generate cluster labels for new input images. This classifier model can then be used to process new images in an efficient manner, and classify them into appropriate clusters.
-
-