Patents by Inventor Miroslav Trajkovic
Miroslav Trajkovic has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 11928550Abstract: Methods and apparatus to locate and decode an arranged plurality of barcodes in an image are disclosed. An example method includes obtaining image data representing an image of an environment appearing within a FOV of an imaging device that includes the image sensor, wherein an arranged plurality of barcodes appear in the image. A first subset of the plurality of barcodes is decoded from the image data. One or more parameters representing a predicted arrangement of the plurality of barcodes in the image is determined based upon location information associated with each of the decoded first subset of the plurality of barcodes. Possible locations for respective ones of a second subset of the plurality of barcodes are determined based upon the one or more parameters, and the second subset of the plurality of barcodes are attempted to be decoded from the image data in vicinities of the respective possible locations.Type: GrantFiled: January 28, 2022Date of Patent: March 12, 2024Assignee: Zebra Technologies CorporationInventors: Peizheng Ma, Eugene B. Joseph, Duanfeng He, Miroslav Trajkovic
-
Publication number: 20240037907Abstract: A method includes: responsive to a scan command, controlling a sensor assembly to scan a machine-readable indicium within a sensor field of view; obtaining an image corresponding to the sensor field of view; determining, from the machine-readable indicium, a decoded item identifier and a scan confidence level associated with the decoded item identifier; determining, from the image, a classified item identifier corresponding to the machine-readable indicium, and a classification confidence level associated with the classified item identifier; selecting, based on the scan confidence level and the classification confidence level, one of the decoded item identifier and the classified item identifier; and generating output data based on the selected one of the decoded item identifier and the classified item identifier.Type: ApplicationFiled: July 27, 2022Publication date: February 1, 2024Inventors: David S. Koch, Miroslav Trajkovic, Yan Zhang, Sam Leitch, Dimitry Kapmar
-
Publication number: 20230356966Abstract: A method includes: receiving, for items in a staging area, item identifiers and staged locations in facility coordinates, as detected via a sensor; obtaining a load configuration including a load sequence defining an order for loading the items into a container, and a load location within the container for each item, in container coordinates; according to the load sequence, for each item: retrieving the corresponding staged location (in facility coordinates) and load location (in container coordinates); providing loading guidance including the retrieved staged location and the retrieved load location, for rendering of the loading guidance in one of the facility coordinate system and the container coordinate system; and responsive to placement of the item in the container, receiving and storing an actual load location, in the container coordinate system; and during an unloading operation, presenting unloading guidance including the actual load locations of the items.Type: ApplicationFiled: May 9, 2022Publication date: November 9, 2023Inventors: Patrenahalli M. Narendra, Miroslav Trajkovic
-
Patent number: 11763412Abstract: Methods and systems to focus an imager for machine vision applications are disclosed. A disclosed example machine vision method includes: capturing, via an imaging assembly, an image of an indicia appearing within a field of view (FOV) of the imaging assembly; recognizing, via a controller, the indicia as a focus adjustment trigger, the focus adjustment trigger operative to trigger an adjustment of at least one focus parameter associated with the imaging assembly; adjusting the at least one focus parameter based at least in part on the indicia; locking the at least one focus parameter such that the at least one focus parameter remains unaltered for a duration; and responsive to the locking of the at least one focus parameter, capturing, via the imaging assembly, at least one subsequent image of an object of interest.Type: GrantFiled: May 12, 2022Date of Patent: September 19, 2023Assignee: Zebra Technologies CorporationInventors: Igor Vinogradov, Miroslav Trajkovic, Heng Zhang
-
Publication number: 20230274820Abstract: A system and method for resource and task allocation are disclosed herein. An example system includes a tracking subsystem configured to track locations and availability of resources; a data acquisition subsystem configured to obtain current status attributes for a patient; a condition detection engine configured to determine, based on the current status attributes and historical medical data, a predicted condition for the patient; a response engine configured to determine, based on the predicted condition, the historical medical data, and a medical history for the patient, a response plan including a resource type to respond to the predicted condition; an allocation engine configured to: obtain the response plan from the response engine; obtain, from the tracking subsystem, tracking data for each resource of the resource type; allocate a resource to be administered to the patient based on the current status attributes of the patient and the tracking data of the resource.Type: ApplicationFiled: February 28, 2022Publication date: August 31, 2023Inventors: Miroslav Trajkovic, Stuart Peter Hubbard
-
Publication number: 20230244891Abstract: Methods and apparatus to locate and decode an arranged plurality of barcodes in an image are disclosed. An example method includes obtaining image data representing an image of an environment appearing within a FOV of an imaging device that includes the image sensor, wherein an arranged plurality of barcodes appear in the image. A first subset of the plurality of barcodes is decoded from the image data. One or more parameters representing a predicted arrangement of the plurality of barcodes in the image is determined based upon location information associated with each of the decoded first subset of the plurality of barcodes. Possible locations for respective ones of a second subset of the plurality of barcodes are determined based upon the one or more parameters, and the second subset of the plurality of barcodes are attempted to be decoded from the image data in vicinities of the respective possible locations.Type: ApplicationFiled: January 28, 2022Publication date: August 3, 2023Inventors: Peizheng Ma, Eugene B. Joseph, Duanfeng He, Miroslav Trajkovic
-
Publication number: 20230199310Abstract: A system and methods for performing focus tuning of an imaging system. The method includes an image sensor obtaining a plurality of images, with each image obtained at a different focus of a tunable optical element. A processor identifying a reference element in at least one of the images of the plurality of images. The processor then determines a reference image that includes the reference element and identifies a reference focus of the tunable optical element. The processor then stores the identified reference focus in a memory.Type: ApplicationFiled: December 17, 2021Publication date: June 22, 2023Inventors: Miroslav Trajkovic, Heng Zhang, Igor Vinogradov
-
Publication number: 20230112116Abstract: An example method includes: during a container load process, controlling a sensor assembly to capture sensor data depicting a container interior; detecting, from the sensor data, items in the container interior; determining, based on the detected items, first and second load process metrics associated with first and second targets; generating first and second normalized metrics based on the first and second load process metrics, and the first and second targets; obtaining a first weighting factor associated with the first load process metric, and a second weighting factor associated with the second load process metric; combining the first normalized metric and the first weighting factor, with the second normalized metric and the second weighting factor, to generate an aggregated load process metric; and transmitting a control command according to the aggregated load process metric; and rendering, at an indicator device, a load process state indicator according to the control command.Type: ApplicationFiled: September 30, 2021Publication date: April 13, 2023Inventors: Adithya H. Krishnamurthy, Miroslav Trajkovic, Justin F. Barish
-
Patent number: 11562561Abstract: Systems and methods for object recognition with limited input are disclosed herein. An example method includes updating a neural network trained to perform object recognition on a first rendition of an object, so that the neural network performs object recognition on a second rendition of the object, using a limited set of input images. The method includes receiving a limited set of model images of the second rendition of the object, accessing a corresponding image mapping, and generating a large number of training images from the limited set, where image mappings include geometric, illumination, and/or obscuration transformations. The neural network is then trained, from this initial small set, to classify the second rendition of the object.Type: GrantFiled: June 7, 2021Date of Patent: January 24, 2023Assignee: Zebra Technologies CorporationInventors: Duanfeng He, Miroslav Trajkovic
-
Patent number: 11443449Abstract: Three-dimensional (3D) imaging systems and methods are described for implementing virtual grading of package walls in commercial trailer loading. The 3D imaging systems and methods comprise capturing, by a 3D-depth camera, 3D image data of a vehicle storage area. A package wall is determined, based on the set of 3D image data and by a 3D data analytics application (app) executing on one or more processors communicatively coupled to the 3D-depth camera, a package wall within the vehicle storage area. A wall grade is assigned to the package wall, where the package wall is defined within the set of 3D image data by a plurality of packages each having a similar depth dimension, and wherein the package wall has a wall height.Type: GrantFiled: February 26, 2020Date of Patent: September 13, 2022Assignee: Zebra Technologies CorporationInventors: Adithya H. Krishnamurthy, Justin F. Barish, Miroslav Trajkovic
-
Publication number: 20220270383Abstract: Methods and systems to focus an imager for machine vision applications are disclosed. A disclosed example machine vision method includes: capturing, via an imaging assembly, an image of an indicia appearing within a field of view (FOV) of the imaging assembly; recognizing, via a controller, the indicia as a focus adjustment trigger, the focus adjustment trigger operative to trigger an adjustment of at least one focus parameter associated with the imaging assembly; adjusting the at least one focus parameter based at least in part on the indicia; locking the at least one focus parameter such that the at least one focus parameter remains unaltered for a duration; and responsive to the locking of the at least one focus parameter, capturing, via the imaging assembly, at least one subsequent image of an object of interest.Type: ApplicationFiled: May 12, 2022Publication date: August 25, 2022Inventors: Igor Vinogradov, Miroslav Trajkovic, Heng Zhang
-
Patent number: 11335105Abstract: Methods and systems to focus an imager for machine vision applications are disclosed. A disclosed example machine vision method includes: capturing, via an imaging assembly, an image of an indicia appearing within a field of view (FOV) of the imaging assembly; recognizing, via a controller, the indicia as a focus adjustment trigger, the focus adjustment trigger operative to trigger an adjustment of at least one focus parameter associated with the imaging assembly; adjusting the at least one focus parameter based at least in part on the indicia; locking the at least one focus parameter such that the at least one focus parameter remains unaltered for a duration; and responsive to the locking of the at least one focus parameter, capturing, via the imaging assembly, at least one subsequent image of an object of interest.Type: GrantFiled: October 8, 2019Date of Patent: May 17, 2022Assignee: Zebra Technologies CorporationInventors: Igor Vinogradov, Miroslav Trajkovic, Heng Zhang
-
Publication number: 20210295043Abstract: Systems and methods for object recognition with limited input are disclosed herein. An example method includes updating a neural network trained to perform object recognition on a first rendition of an object, so that the neural network performs object recognition on a second rendition of the object, using a limited set of input images. The method includes receiving a limited set of model images of the second rendition of the object, accessing a corresponding image mapping, and generating a large number of training images from the limited set, where image mappings include geometric, illumination, and/or obscuration transformations. The neural network is then trained, from this initial small set, to classify the second rendition of the object.Type: ApplicationFiled: June 7, 2021Publication date: September 23, 2021Inventors: Duanfeng He, Miroslav Trajkovic
-
Publication number: 20210264630Abstract: Three-dimensional (3D) imaging systems and methods are described for implementing virtual grading of package walls in commercial trailer loading. The 3D imaging systems and methods comprise capturing, by a 3D-depth camera, 3D image data of a vehicle storage area. A package wall is determined, based on the set of 3D image data and by a 3D data analytics application (app) executing on one or more processors communicatively coupled to the 3D-depth camera, a package wall within the vehicle storage area. A wall grade is assigned to the package wall, where the package wall is defined within the set of 3D image data by a plurality of packages each having a similar depth dimension, and wherein the package wall has a wall height.Type: ApplicationFiled: February 26, 2020Publication date: August 26, 2021Inventors: Adithya H. Krishnamurthy, Justin F. Barish, Miroslav Trajkovic
-
Patent number: 11062104Abstract: A barcode reader is provided. The barcode reader includes a first image acquisition assembly having a first imager assembly configured to capture infrared (IR) light and an IR illumination assembly configured to emit IR light over at least a portion of a first field of view (FOV) of the first imager assembly so as to illuminate targets within the first FOV. The barcode reader further includes a second image acquisition assembly having a second imager assembly configured to capture visible light and a visible-light illumination assembly configured to emit visible light over at least a portion of a second FOV of the second imager assembly so as to illuminate targets within the second FOV.Type: GrantFiled: July 8, 2019Date of Patent: July 13, 2021Assignee: Zebra Technologies CorporationInventors: Duanfeng He, Miroslav Trajkovic
-
Patent number: 11062138Abstract: Systems and methods for object recognition with limited input are disclosed herein. An example method includes updating a neural network trained to perform object recognition on a first rendition of an object, so that the neural network performs object recognition on a second rendition of the object, using a limited set of input images. The method includes receiving a limited set of model images of the second rendition of the object, accessing a corresponding image mapping, and generating a large number of training images from the limited set, where image mappings include geometric, illumination, and/or obscuration transformations. The neural network is then trained, from this initial small set, to classify the second rendition of the object.Type: GrantFiled: November 19, 2019Date of Patent: July 13, 2021Assignee: Zebra Technologies CorporationInventors: Duanfeng He, Miroslav Trajkovic
-
Publication number: 20210150210Abstract: Systems and methods for object recognition with limited input are disclosed herein. An example method includes updating a neural network trained to perform object recognition on a first rendition of an object, so that the neural network performs object recognition on a second rendition of the object, using a limited set of input images. The method includes receiving a limited set of model images of the second rendition of the object, accessing a corresponding image mapping, and generating a large number of training images from the limited set, where image mappings include geometric, illumination, and/or obscuration transformations. The neural network is then trained, from this initial small set, to classify the second rendition of the object.Type: ApplicationFiled: November 19, 2019Publication date: May 20, 2021Inventors: Duanfeng He, Miroslav Trajkovic
-
Patent number: 11003804Abstract: Embodiments of the present invention are generally directed to system and methods for estimating the time associated with completion of loading and/or unloading of a container. In an embodiment, the present invention is a method of estimating an estimated time to completion (ETC) of loading a container. The method includes: capturing, via an image capture apparatus, a three-dimensional image representative of a three-dimensional formation, the three-dimensional image having a plurality of points with three-dimensional point data; based at least in part on a first sub-plurality of the points, determining an active load time for the container; based at least in part on a second sub-plurality of the points, determining a fullness of the container; and estimating, by a controller, the ETC based on the active load time and on the fullness.Type: GrantFiled: December 22, 2017Date of Patent: May 11, 2021Assignee: Symbol Technologies, LLCInventors: Adithya H. Krishnamurthy, Miroslav Trajkovic
-
Patent number: 10977813Abstract: A system and method for performing robust depth calculations with time of flight (ToF) sensors using multiple exposure times is disclosed. A three-dimensional (3D) depth sensor assembly captures a first array of n point values, where each point value of the first array has a respective first-array depth component and a respective first-array quality component. The 3D depth sensor assembly then captures a second array of n point values, where each point value of the second array has a respective second-array depth component and a respective second-array quality component. A processor then renders a 3D point cloud comprising a third array of n point values, where each point value of the third array has a respective third-array depth component. The respective third-array depth component for each point value of the third array is based on either the corresponding respective first-array depth component or the corresponding respective second-array depth component.Type: GrantFiled: August 9, 2019Date of Patent: April 13, 2021Assignee: Zebra Technologies CorporationInventors: Miroslav Trajkovic, Adithya H. Krishnamurthy, Bryn P. Martino
-
Patent number: 10977514Abstract: A method and apparatus for template matching to find a predetermined pattern in an image is disclosed. A first visual boundary is detected in the captured image, and a second boundary concentric with the first boundary is calculated. The first and second boundaries define a portion of the captured image. The portion of the captured image is incrementally scanned about the center of the second boundary for a predetermined pattern having a predetermined orientation within the portion that match a template image. Alternatively, the portion of the captured image is unwrapped into a linear band image such that the first and second boundaries form a linear top and linear bottom of the linear band image, and the linear band image of the portion of the captured image is scanned for a predetermined pattern that matches a template image.Type: GrantFiled: November 16, 2018Date of Patent: April 13, 2021Assignee: Zebra Technologies CorporationInventors: Miroslav Trajkovic, Heng Zhang