公开(公告)号：US12026226B2

公开(公告)日：2024-07-02

申请号：US17408674

申请日：2021-08-23

Applicant: CARNEGIE MELLON UNIVERSITY

Inventor： Marios Savvides ,  Chenchen Zhu ,  Fangyi Chen ,  Uzair Ahmed ,  Ran Tao

IPC: G06F18/2136 ,  G06N3/04 ,  G06N5/02

CPC classification number: G06F18/2136 ,  G06N3/04 ,  G06N5/02

Abstract: Disclosed herein is an improved few-shot detector which utilizes semantic relation reasoning to learn novel objects from both visual information and the semantic relation of base class objects Specifically, a semantic space is constructed using word embeddings. Guided by the word embeddings of the classes, the detector is trained to project the objects from the visual space to the semantic space and to align their image representations with the corresponding class embeddings.

公开(公告)号：US20240054764A1

公开(公告)日：2024-02-15

申请号：US18267543

申请日：2022-02-04

Applicant: CARNEGIE MELLON UNIVERSITY

Inventor： Ran Tao ,  Marios Savvides ,  Han Zhang

IPC: G06V10/771

CPC classification number: G06V10/771

Abstract: Disclosed herein is a methodology implementing feature map-level data augmentation in a feature map. Two or more units in the feature map are selected and the values of locations in the two or more units are swapped among the two or more units. Value perturbations applied around local units in the feature map implicitly lead to an unused data augmentation at the image level.

公开(公告)号：US20250045692A1

公开(公告)日：2025-02-06

申请号：US18808723

申请日：2024-08-19

Applicant: CARNEGIE MELLON UNIVERSITY

Inventor： Marios Savvides ,  Sreena Nallamothu ,  Magesh Kannan ,  Uzair Ahmed ,  Ran Tao ,  Yutong Zheng

IPC: G06Q10/087 ,  G06F16/583 ,  G06F18/2113 ,  G06F18/214 ,  G06F18/28 ,  G06V10/40 ,  G06V10/75 ,  G06V20/52 ,  H04N7/18

Abstract: Disclosed herein is a system and method of identifying misplaced products on a retail shelf using a feature extractor trained to extract features from images of products on the shelf and output identifying information regarding the product in the product image. The extracted features are compared to extracted features in a product library and a best fit is obtained. A misplaced product is identified if the identifying information produced by the feature extractor fails to match the identifying information associated with the best fit features from the product library.

公开(公告)号：US12190565B2

公开(公告)日：2025-01-07

申请号：US18267540

申请日：2022-02-03

Applicant: CARNEGIE MELLON UNIVERSITY

Inventor： Ran Tao ,  Marios Savvides

IPC: G06K9/00 ,  G06V10/764 ,  G06V10/77 ,  G06V10/771

Abstract: Disclosed herein is a methodology for refining novel-class features in a few-shot learning scenario by fine-tuning the feature extractor by reducing both class-agnostic biases and class-specific biases. A distribution calibration module is used to reduce the class-agnostic bias by normalizing the overall feature distribution for novel classes and further reshaping the feature manifold for fast adaptation during fine-tuning. Selected sampling is used to reduce class-specific bias by augmenting more data for better estimation.

公开(公告)号：US20240355085A1

公开(公告)日：2024-10-24

申请号：US18587200

申请日：2024-02-26

Applicant: CARNEGIE MELLON UNIVERSITY

Inventor： Marios Savvides ,  Chenchen Zhu ,  Fangyi Chen ,  Uzair Ahmed ,  Ran Tao

IPC: G06V10/44 ,  G06F18/214 ,  G06T7/73 ,  G06V10/25 ,  G06V20/20

CPC classification number: G06V10/443 ,  G06F18/214 ,  G06T7/73 ,  G06V10/25 ,  G06V20/20

Abstract: Disclosed herein is a system and method for matching products detected in an image of a shelf. The match or non-match of the products is then used to make a determination that the products are correctly positioned on the shelf of if the positioning of the products represents a plug or spread situation.

公开(公告)号：US12067527B2

公开(公告)日：2024-08-20

申请号：US17400996

申请日：2021-08-12

Applicant: CARNEGIE MELLON UNIVERSITY

Inventor： Marios Savvides ,  Sreena Nallamothu ,  Magesh Kannan ,  Uzair Ahmed ,  Ran Tao ,  Yutong Zheng

IPC: G06Q10/087 ,  G06F16/583 ,  G06F18/2113 ,  G06F18/214 ,  G06F18/28 ,  G06V10/40 ,  G06V10/75 ,  G06V20/52 ,  H04N7/18

CPC classification number: G06Q10/087 ,  G06F16/5846 ,  G06F18/2113 ,  G06F18/214 ,  G06F18/28 ,  G06V10/40 ,  G06V10/751 ,  G06V20/52 ,  H04N7/18

Abstract: Disclosed herein is a system and method of identifying misplaced products on a retail shelf using a feature extractor trained to extract features from images of products on the shelf and output identifying information regarding the product in the product image. The extracted features are compared to extracted features in a product library and a best fit is obtained. A misplaced product is identified if the identifying information produced by the feature extractor fails to match the identifying information associated with the best fit features from the product library.

公开(公告)号：US11915463B2

公开(公告)日：2024-02-27

申请号：US17408778

申请日：2021-08-23

Applicant: CARNEGIE MELLON UNIVERSITY

Inventor： Marios Savvides ,  Chenchen Zhu ,  Fangyi Chen ,  Uzair Ahmed ,  Ran Tao

IPC: G06V10/44 ,  G06T7/73 ,  G06V10/25 ,  G06V20/20 ,  G06F18/214

CPC classification number: G06V10/443 ,  G06F18/214 ,  G06T7/73 ,  G06V10/25 ,  G06V20/20

Abstract: Disclosed herein is a system and method of identifying new products on a retail shelf using a feature extractor trained to extract features from images of products on the shelf and output identifying information regarding the product in the product image. The extracted features are compared to extracted features in a product library and a best-fit is obtained. A new product is identified if the distance between the features of the product on the shelf and the features of the best-fit product from the product library are above a predetermined threshold.

公开(公告)号：US20220058432A1

公开(公告)日：2022-02-24

申请号：US17408674

申请日：2021-08-23

Applicant: CARNEGIE MELLON UNIVERSITY

Inventor： Marios Savvides ,  Chenchen Zhu ,  Fangyi Chen ,  Uzair Ahmed ,  Ran Tao

IPC: G06K9/62 ,  G06N5/02 ,  G06N3/04

Abstract: Disclosed herein is an improved few-shot detector which utilizes semantic relation reasoning to learn novel objects from both visual information and the semantic relation of base class objects Specifically, a semantic space is constructed using word embeddings. Guided by the word embeddings of the classes, the detector is trained to project the objects from the visual space to the semantic space and to align their image representations with the corresponding class embeddings.

公开(公告)号：US12189714B2

公开(公告)日：2025-01-07

申请号：US18266744

申请日：2022-02-02

Applicant: CARNEGIE MELLON UNIVERSITY

Inventor： Marios Savvides ,  Chenchen Zhu ,  Fangyi Chen ,  Uzair Ahmed ,  Ran Tao

IPC: G06F18/2136 ,  G06N3/04 ,  G06N5/02

Abstract: Disclosed herein is an improved few-shot detector which utilizes a dynamic semantic network which takes as input a language feature and generates trainable parameters for a visual network. The visual network takes a visual feature as input and generates a classification and localization of an object.

公开(公告)号：US20240071050A1

公开(公告)日：2024-02-29

申请号：US18267540

申请日：2022-02-03

Applicant: CARNEGIE MELLON UNIVERSITY

Inventor： Ran Tao ,  Marios Savvides

IPC: G06V10/77 ,  G06V10/764 ,  G06V10/771

CPC classification number: G06V10/7715 ,  G06V10/764 ,  G06V10/771

Abstract: Disclosed herein is a methodology for refining novel-class features in a few-shot learning scenario by fine-tuning the feature extractor by reducing both class-agnostic biases and class-specific biases. A distribution calibration module is used to reduce the class-agnostic bias by normalizing the overall feature distribution for novel classes and further reshaping the feature manifold for fast adaptation during fine-tuning. Selected sampling is used to reduce class-specific bias by augmenting more data for better estimation.