Abstract: Disclosed is a modeling method for a picking target of a fruit bunch picking robot, which relates to the technical field of general image data processing or generation. The modeling method includes: obtaining an image of a to-be-picked region of a picking robot, and extracting image features of each branch and fruit cluster in the image of the to-be-picked region through a multi-task perception network; determining a to-be-picked fruit cluster based on the image feature of the fruit cluster; inputting the image features of the branch and the fruit cluster into a subordinate decision model to determine a branch connected to the to-be-picked fruit cluster; and extracting key points of image features of the to-be-picked fruit cluster and the branch connected to the to-be-picked fruit cluster, and modeling the to-be-picked fruit cluster and the branch connected to the to-be-picked fruit cluster.

Abstract: Provided are a method and device for annotating a detected object. The method of annotating a detected object may include classifying a class of the object, generating a bounding box including a first quadrangle and a second quadrangle both sharing one side, for the object based on the class of the object, and generating an annotation on the object in units of the bounding box.

Abstract: A method for cable wire insertion monitoring includes receiving, from a camera system, an inspection image of a connector housing held in a housing retainer of a housing inspection system. Insertion of a cable wire into a corresponding cable cavity of the connector housing is detected via an insertion monitoring machine vision system. Based at least in part on the inspection image, an insertion depth of the cable wire into the corresponding cable cavity is estimated. An indication of the insertion depth of the cable wire is output.

Abstract: A system and method to detect the positioning of a person relative to a seat includes an image sensor and a processing system. The image sensor is directed toward the seat and is configured to capture images of the person and the seat and to supply image data indicative of the captured images. The processing system is in operable communication with the image sensor and is configured to receive and process the image data to determine when the person is not in a first position by determining that the image sensor is not focused on the person for a first predetermined time period, based on which of a plurality of objects on the seat are visible in the captured images, and/or by variations in the illumination pattern emitted onto the person.

Abstract: An information processing apparatus according to an embodiment includes one or more hardware processors configured to function as a correspondence relationship acquisition unit, a selection unit, and a calculation unit. The correspondence relationship acquisition unit is configured to calculate a plurality of image features from a plurality of images captured by a camera and acquire a correspondence relationship between the image features. The selection unit is configured to select a plurality of correspondence relationships, based on effectiveness of the correspondence relationship and an influence on at least one of the images when the correspondence relationship with the effectiveness lower than an effectiveness threshold is eliminated. The calculation unit is configured to calculate at least one of a position and direction of the camera and three-dimensional information on the image features from the correspondence relationship selected from among the correspondence relationships.

Abstract: There are provided systems and methods comprising, for each given overlay target of a plurality of different overlay targets to be manufactured on a semiconductor specimen, said given overlay target comprising a plurality of stacked semiconductor layers, obtaining a design image of the given overlay target, feeding the design image to a trained machine learning model, to simulate at least one image of the given overlay target that would have been acquired by an electron beam examination system, using the at least one image to determine, before actual manufacturing of the given overlay target, data informative of at least one simulated overlay in the image, and using the data informative of the at least one simulated overlay of each given overlay target to select at least one optimal overlay target among the plurality of different overlay targets, the optimal overlay target being usable to be manufactured on the semiconductor specimen.

Abstract: Disclosed embodiments may include a method for dynamically detecting disabilities. The method may include receiving image data and sensory data from an imaging device and one or more sensors of a user interacting with a machine, determining, using a machine learning model, whether the user interacting with the machine is impaired based on the image data or the sensory data, or combinations thereof, and in response to determining that the user is impaired: causing the machine to output a sound or one or more vibration sequences within a predetermined distance from a reader of the machine, initiating a phone call to a user device associated with the user using a phone number associated with the data, initiate a machine process, receiving, from the user device, a command to terminate the machine process, and transmitting a termination instruction to the machine.

Abstract: Systems and techniques are described herein for tracking objects. For instance, a method for tracking objects is provided. The method may include generating an output embedding based on an object in a first image; obtaining a predicted position of the object in a second image; modifying the output embedding based on the predicted position to generate a modified output embedding; and detecting the object in the second image based on the modified output embedding.

Abstract: Systems and techniques that facilitate video behavior recognition and prediction based on a generative language model are provided. One or more embodiments described herein can comprise a system, which can comprise a processor set, a set of one or more computer-readable storage media, and program instructions collectively stored in the set of one or more computer-readable storage media. The program instructions can cause the processor set to input a video to a first machine learning model to produce text data that describes the video, input the text data into a second machine learning model to produce a first set of predictions regarding one or more agents depicted in the video, input the first set of predictions into an embedding layer to produce embeddings, and input the embeddings and the video into a third machine learning model to produce a second set of predictions regarding the one or more agents.

Abstract: The present application discloses a low-light image enhancement method based on wavelet transform and Retinex-Net, and belongs to the field of image enhancement technology. In order to solve the problems of colour distortion and lack of details such as edges and textures when Retinex-Net processes some images, the present application adds the wavelet transform, fuses the high-frequency components with regional characteristics, and processes only the low-frequency component with Retinex-Net, to retain more details of edges and textures in the image, and transfers the low-frequency and high-frequency components from the HSV space to the value space for value fusion and stretching, to effectively improve the colour distortion of the image, and also improve the expression of details of the image. The present application retains more detail information while avoiding colour distortion, and the enhancement effect is satisfactory.

Abstract: Presented herein are systems and methods for the employment of machine learning models for image processing as may be performed by computing devices associated with an end user. A method may include obtaining video data comprising a plurality of frames including a document of a document type. The method may include executing an object recognition engine of a machine-learning architecture using image data of the plurality of frames, the object recognition engine trained to detect edges of documents. The method may include identifying, based on the edge detection, a plurality of boundaries for the document. The method may include validating, based on the plurality of boundaries, the document as the document type. The method may include transmitting via one or more networks, to a computer remote from the computing device, responsive to the validation of the type of document, the image data for the plurality of frames depicting the document.

Abstract: The present invention provides a learning apparatus (10) including an acquisition unit (11) that acquires learning data associating a training image including a person with a correct answer label indicating a position of each person, a correct answer label indicating whether each of a plurality of keypoints of a body of each of the persons is visible in the training image, and a correct answer label indicating a position, within the training image, of the keypoint being visible in the training image among a plurality of the keypoints, and a learning unit (12) that learns, based on the learning data, an estimation model.

Abstract: Disclosed are a device and method with pose estimation. A method of operating an electronic device includes: obtaining monocular images captured at different respective points in time by a same camera; estimating, based on a neural network, a position change of the camera corresponding to the time between the points in time at which the images were captured; and estimating a position and/or orientation of the camera based on the estimated position change of the camera and based on the images.

Abstract: There is provided a system of training a deep learning model for support of a device in a remote technical support system, the system comprising a processing circuitry configured to: receive a plurality of images of a device to be supported; provide a user interface of annotating each image of the plurality of images, the annotating being data indicative of one or more of: one or more image identification labels, and one or more image subcomponent label tuples; and utilize, at least, the plurality of annotated images, to train a deep learning model to classify and/or segment an image to data indicative of at least one of: one or more image identification labels, and one or more image subcomponent label tuples, thereby providing a platform enabling automated creation of a remote support service.

Abstract: Presented herein are systems and methods for the employment of machine learning models for image processing. A method may include a capture of a video feed including image data of a document at a client device. The client device can provide the video feed to another computing device. The method can include, by the client device or the other computing device object recognition for recognizing a type of document and capturing an image exceeding a quality threshold of the document amongst the frames within the video feed. The method may further include the execution of other image processing operations on the image data to improve the quality of the image or features extracted therefrom. The method may further include anti-fraud detection or scoring operations to determine an amount of risk associated with the image data.

Abstract: A distance measuring device includes a first imaging element having a first pixel having a plurality of photoelectric conversion regions, a second imaging element having a second pixel and having a first parallax in a first direction between the second imaging element and the first imaging element, and distance measuring means for generating integrated distance measuring information based on signals output from the first imaging element and the second imaging element, and the plurality of photoelectric conversion regions have a second parallax in a second direction intersecting the first direction.

Abstract: In an evaluation method, first, spheroids obtained by three-dimensional culture of multiple kinds of liver-derived cells are imaged by optical coherence photography, a localization region is extracted from the photographic image, the localization region is analyzed (analysis step), and the condition of the spheroid is evaluated. The analysis step includes a first calculation step of calculating the area of the entire spheroid in the photographic image, a second calculation step of calculating the area of the localization region, a third calculation step of calculating the ratio of the localization region on the basis of the two areas, and a fourth calculation step of calculating an evaluation parameter on the basis of the ratio.

Abstract: An information processing apparatus according to the present disclosure includes: a projection unit configured to arrange a predetermined number of virtual points in a front direction of an imaging device that captures a first image, by using an estimated position of the imaging device as a starting point, and project the virtual points onto each of a plurality of images used for generating an environment map being three-dimensional information; an image extraction unit configured to extract a second image from the plurality of images, based on a result of the projection of the virtual points; and an estimation unit configured to estimate, by using a feature point included in a target image being a target for estimating a position and a pose of an imaging device and a feature point included in the second image, aa position and a pose of the imaging device that captures the target image.

Abstract: Provided is a motion data generation device including an acquisition unit that acquires a plurality of pieces of motion data to be data converted, a conversion data choosing unit that groups the plurality of pieces of motion data for each motion class that is a target motion for data augmentation, a data conversion unit that sets at least one piece of the motion data grouped for the each motion class for reference data, sets at least one piece of the motion data different from the reference data among the grouped motion data for data to be converted, and generates extension motion data in which the data to be converted is synchronized with reference to motion timing of the reference data, and an output unit that outputs the generated extension motion data.

Abstract: Disclosed herein are apparatuses and methods for auto-configuring a region of interest (ROI) associated with a camera. In one implementation, A method includes tracking, over a plurality of image frames spanning a first period of time, a plurality of objects moving in an environment, and determining a plurality of respective trajectories of the plurality of objects. The method includes identifying a common trajectory shared by more than one object of the plurality of objects based on comparing each of the plurality of respective trajectories to one another and identifying an ROI that encompasses the common trajectory. The method further includes tracking, over a plurality of new image frames spanning a second period of time, another plurality of objects moving in the environment, and generating an alert that the common trajectory has changed based on new trajectories of the another plurality of objects.