公开(公告)号：WO2023057494A1

公开(公告)日：2023-04-13

申请号：PCT/EP2022/077660

申请日：2022-10-05

Applicant: SIEMENS AG ÖSTERREICH

Inventor： HIESSL, Thomas ,  REZAPOUR LAKANI, Safoura ,  SCHALL, Daniel

IPC: G06N3/08 ,  G06N20/00 ,  G06V10/82 ,  G06F18/2433 ,  G06N3/0455 ,  G06N3/098 ,  G06V10/95 ,  G06V20/60 ,  G06V2201/06

Abstract: Computer-implementiertes Verfahren zur Anomalie-Erkennung in Sensordaten (SD1, SD2), umfassend folgende Schritte: a) Erzeugen und Trainieren eines ersten und zumindest eines zweiten lokalen Modells jeweils auf Basis eines Autoen- coders und Bestimmen eines lokalen Schwellwerts (TH1, TH2) für die jeweilige lokale Modell-Ausgangsgröße (RD), b) Übermitteln der lokalen Modellgewichte (W1, W2) und der lokalen Schwellwerte (TH1, TH2 ) an einen Server (S), c) Erzeugen und Trainieren eines globalen Modells (GM) auf Basis eines Autoencoders und Bestimmen eines globalen Schwellwerts (TH GM) für die globale Modell- Ausgangsgröße (RD, W_GM), ti) Übermitteln der globalen Modellgewichte (W_GM) und des globalen Schwellwerts (TH_GM) an den ersten Klien- ten (C1), und Übernehmen der globalen Modellgewich- te (W_GM) für das erste lokale Modell (LM1), e) Erfassen von ersten Sensordaten (SD1) durch ein erstes Sensormittel (SM1), f) Anwenden der ersten Sensordaten (SD1) auf das erste loka- le Modell (LM1) und Bestimmen der lokalen Modell- Ausgangsgröße (RD, LM1_OUT) des ersten Klienten (C1), g) Erkennen einer Anomalie (ANO) für die Sensordaten, falls die lokale Modell-Ausgangsgröße (RD, LM1_OUT) außerhalb eines Bereichs liegt, welcher durch den globalen Schwellwert (TH_GM) festgelegt ist.

公开(公告)号：WO2023035079A1

公开(公告)日：2023-03-16

申请号：PCT/CA2022/051354

申请日：2022-09-09

Applicant: BRIGHTER SIGHT INC.

Inventor： CHAWLA, Bobby ,  AMIRABADI, Sepehr Eslami

IPC: G16H20/17 ,  A61M5/315 ,  G06V10/82 ,  G06V20/60

Abstract: A system and method for providing information on medication dosage in a syringe prior to administration to a patient comprises determining, from an image of the syringe containing the medication to be administered, a capacity of the syringe; locating, from the image of the syringe, plunger and barrel heads of the syringe; and estimating a dosage volume of the medication in the syringe using a mathematical model based on the capacity of the syringe and using the located plunger and barrel heads as inputs for the mathematical model.

公开(公告)号：WO2023085992A1

公开(公告)日：2023-05-19

申请号：PCT/SE2022/050968

申请日：2022-10-24

Applicant: OPTICEPT TECHNOLOGIES AB

Inventor： DEMIR WESTMAN, Eda ,  SCHMIDT, John ,  DYMEK KRAKOWIAK, Katarzyna Maria ,  CARENBORN, Hanshenric

IPC: G06T7/00 ,  A01H5/00 ,  G06V10/40 ,  G06V10/764 ,  G06V20/60

Abstract: The present invention describes method for analyzing one or more cut flowers with reference to their life length, said method comprising: - taking images of said one or more cut flowers; and - analyzing the images with a software tool, involving labelling said one or more cut flowers into at least two quality categories.

公开(公告)号：WO2023082018A1

公开(公告)日：2023-05-19

申请号：PCT/CA2022/051676

申请日：2022-11-14

Applicant: TECHINSIGHTS INC.

Inventor： PAWLOWICZ, Christopher ,  GREEN, Michael ,  MACHADO TRINDADE, Bruno ,  PULLELA, Sneha ,  YU, Zifan ,  REN, Fengbo

IPC: G06V20/60 ,  G06N3/08 ,  G06N3/0464 ,  G06V10/26 ,  G06V10/82

Abstract: Described are various embodiments of a machine learning system and method for object-specific recognition.

公开(公告)号：WO2023061599A1

公开(公告)日：2023-04-20

申请号：PCT/EP2021/078505

申请日：2021-10-14

Applicant: SES-IMAGOTAG SA

Inventor： GADOU, Thierry ,  SCHWARZ, Thomas ,  RÖSSL, Andreas

IPC: G06Q10/08 ,  G06Q30/06 ,  G09F3/20 ,  G06V20/60

Abstract: Verfahren zum logischen Verknüpfen einer elektronischen Anzeigeeinheit mit einem Produkt durch eine Verknüpfungseinrichtung, wobei sich die Anzeigeeinheit und das Produkt an einer Befestigungsstruktur im Erfassungsbereich einer Bilderfassungseinrichtung befinden, wobei das Verfahren die nachfolgend angeführten Verfahrensschritte aufweist, nämlich: Generieren von Bilddaten, die eine optische Abbildung des Produkts und der Anzeigeeinheit repräsentieren, durch die Bilderfassungseinrichtung und übertragen dieser Bilddaten an die Verknüpfungseinrichtung, mit deren Hilfe die logische Verknüpfung erstellbar ist, Generieren eines Auslöse-Signals durch eine Auslösestufe, die der Befestigungsstruktur zugeordnet ist, an der die Anzeigeeinheit befestigt ist, wobei das Auslöse-Signal das Vorhandensein einer nicht verknüpften Anzeigeeinheit im Erfassungsbereich angibt, und logisches Verknüpfen der noch nicht verknüpften elektronischen Anzeigeeinheit mit dem Produkt unter Beachtung des Auftretens des Auslöse-Signals.

公开(公告)号：WO2023288209A1

公开(公告)日：2023-01-19

申请号：PCT/US2022/073625

申请日：2022-07-12

Applicant: SENSORMATIC ELECTRONICS, LLC

Inventor： SUBRAMANIAN, Gopi ,  CELI, Joseph

IPC: G06V20/52 ,  G06V20/60 ,  G08B17/00 ,  G08B17/12

Abstract: Disclosed herein are apparatuses and methods for providing contextual data for a security event in an environment. An implementation may comprise receiving and parsing sensor data from a plurality of sensors located in the environment. The implementation may comprise storing, in a database, the parsed sensor data comprising identifiers of the plurality of objects and the attributes. The implementation may comprise detecting a security event at the environment and determining a type, a time window, and a location of the security event. The implementation may comprise retrieving, from the database, a set of object identifiers with timestamps within the time window and a matching location of the security event. The implementation may comprise filtering, from the set of object identifiers, at least one object identifier that matches the type of the security event, and outputting the at least one object identifier and attributes of the at least one object identifier.

公开(公告)号：WO2022226225A2

公开(公告)日：2022-10-27

申请号：PCT/US2022/025819

申请日：2022-04-21

Applicant: MAPLEBEAR INC. (DBA INSTACART)

Inventor： YANG, Shiyuan ,  CHANDRA, Shray

IPC: G06V10/40 ,  G01G19/4144 ,  G06Q20/18 ,  G06Q20/208 ,  G06T2200/24 ,  G06T2207/20081 ,  G06T2207/20084 ,  G06T7/10 ,  G06T7/50 ,  G06V10/16 ,  G06V10/23 ,  G06V10/70 ,  G06V10/945 ,  G06V20/60 ,  G06V20/64 ,  G07G1/0036 ,  H04N5/247

Abstract: An item recognition system uses a top camera and one or more peripheral cameras to identify items. The item recognition system may use image embeddings generated based on images captured by the cameras to generate a concatenated embedding that describes an item depicted in the image. The item recognition system may compare the concatenated embedding to reference embeddings to identify the item. Furthermore, the item recognition system may detect when items are overlapping in an image. For example, the item recognition system may apply an overlap detection model to a top image and a pixel-wise mask for the top image to detect whether an item is overlapping with another in the top image. The item recognition system notifies a user of the overlap if detected.

公开(公告)号：WO2022204800A1

公开(公告)日：2022-10-06

申请号：PCT/CA2022/050466

申请日：2022-03-29

Applicant: TERRAMERA, INC.

Inventor： SUTTON, Darren Bryce ,  HAMARNEH, Ghassan ,  PARTIDA, Carolina

IPC: G06V20/60 ,  G06N3/08 ,  G06N20/00 ,  G06V10/764

Abstract: Systems and methods for training a machine learning model for generating a structural representation of a plant are provided, as well as systems and methods for generating a structural representation of a plant via such a model. The training method involves encoding a plant image into a structural representation of the plant (e.g. a "skeleton"), decoding the structural representation of the plant into a reconstructed image of the plant, and classifying the reconstructed image as having been generated based on a ground-truth structural representation or output of the encoder. Such classification incentivizes the encoder to produce structural representations which do not "smuggle" texture information (e.g. appearance, such as color). Texture information may be separately represented. The encoder, once trained, may be used to generate structural representations from plant images without necessarily requiring decoding or classification.

公开(公告)号：WO2022165211A1

公开(公告)日：2022-08-04

申请号：PCT/US2022/014360

申请日：2022-01-28

Applicant: INDIE SEMICONDUCTOR, INC. ,  TRAA, Johannes ,  SCHWEITZER, Andrew M. ,  YELLEPEDDI, Atulya

Inventor： TRAA, Johannes ,  SCHWEITZER, Andrew M. ,  YELLEPEDDI, Atulya

IPC: G01S13/00 ,  G06V20/58 ,  G06V20/60

Abstract: A method includes an operation to collect radar signals reflected from objects in a field of view. Range-angle-doppler bins representing three-dimensional objects in the field of view are formed. A local median operation is used across a selected dimension of the range-angle-doppler bins to eliminate background noise in the range-angle-doppler bins. Low energy peak regions are masked by removing radial velocity values in the selected dimension to form a sparse range-angle two-dimensional grid. The radar signals reflected from objects in the field of view are processed to extract reflection point detections. Reflection point detections are tracked in accordance with short-term filter rules to form tracked reflection point detections. The tracked reflection point detections are formed into clusters. The clusters are processed with long-term filter rules.

公开(公告)号：WO2022132513A1

公开(公告)日：2022-06-23

申请号：PCT/US2021/062301

申请日：2021-12-08

Applicant: CATERPILLAR INC.

Inventor： REAUME, Daniel Joseph ,  RADAKOVIC, Daniela ,  ORGAN, Daniel Jude

IPC: G06V10/82 ,  G06V20/60 ,  G06V10/74

Abstract: A method for object identification using multiple images (100). The method includes training an object identification model (102). Training the model includes collecting training images for each of a plurality of objects (202), labeling each of the plurality of training images with a corresponding one of a plurality of object identifiers (206), and training a neural network with the plurality of labeled training images (208). At least two target images of a target object are received and fed into the trained object identification model (104). The method further includes receiving, from the trained object identification model, for each of the at least two target images, an object identifier corresponding to the target object and a probability that the object identifier corresponds to the target object (106). A similarity value between the at least two target images is computed (108) and the probabilities for the at least two target images are combined in proportion to the similarity value (110).