Abstract: The present invention relates to an image encoding and decoding technique, and more particularly, to an image encoder and decoder using unidirectional prediction. The image encoder includes a dividing unit to divide a macro block into a plurality of sub-blocks, a unidirectional application determining unit to determine whether an identical prediction mode is applied to each of the plurality of sub-blocks, and a prediction mode determining unit to determine a prediction mode with respect to each of the plurality of sub-blocks based on a determined result of the unidirectional application determining unit.

Abstract: An object of the present disclosure is to, in a case of adding a double-page spread in editing of a plurality of albums, determine a criterion for creating the double-page spread to be added to automatically create and add the double-page spread. A program is provided which causes a computer to function as an editing unit configured to edit layouts of a plurality of albums each having an attribute of a common double-page spread common to all of the albums and an attribute of an individual double-page spread in each of the albums as double-page spread attributes. The editing unit automatically creates a double-page spread based on one or more criteria corresponding to the attributes, and adds the created double-page spread to one album among the plurality of album.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods that implement a dual-branched neural network architecture to harmonize composite images. For example, in one or more implementations, the transformer-based harmonization system uses a convolutional branch and a transformer branch to generate a harmonized composite image based on an input composite image and a corresponding segmentation mask. More particularly, the convolutional branch comprises a series of convolutional neural network layers followed by a style normalization layer to extract localized information from the input composite image. Further, the transformer branch comprises a series of transformer neural network layers to extract global information based on different resolutions of the input composite image.

Abstract: Systems and methods to determine an activity associated with an object of interest are presented. The method includes receiving a sequence of a plurality of frames that capture an area of interest during a threshold period of time, wherein an object of interest is present in the area of interest during the threshold period of time. The method also includes analyzing the plurality of frames to detect the object of interest. The method also includes tracking the object of interest in the area of interest during the threshold period of time. The method also includes extracting data indicative of the object of interest. The method also includes predicting, based on the data indicative of the object of interest, an activity associated with the object of interest.

Abstract: The present invention relates to a technique for encoding and decoding video data, and more particularly, to a method for performing inter-prediction in an effective manner. The present invention combines an inter-prediction method using an AMVP mode and an inter-prediction method using a merge mode so as to propose a method for using the same candidate. The method for encoding video data proposed by the present invention comprises the following steps: receiving mode information on an inter-prediction method of a current block; determining, on the basis of the received mode information, whether the interprediction method to be applied to the current block is an AMVP mode or a merge mode; and selecting a candidate to derive motion information of the current block, wherein the candidate is selected in a left region, top region and corner region of the current block and in the same position block as the current block, and the AMVP mode and the merge mode are applied on the basis of the selected candidate.

Abstract: A method for the reliable and error-robust analysis of visually captured vehicle surroundings is provided. A deviation of a camera orientation caused by a force impacting an exterior mirror carrying the camera can be compensated fully automatically. As a result, the method can be used during active travels in the field and creates more safety, in particular for autonomous driving.

Abstract: Multi-task deep learning method for a neural network for automatic pathology detection, comprising the steps: receiving first image data (I) for a first image recognition task; receiving (S2) second image data (V) for a second image recognition task; wherein the first image data (I) is of a first datatype and the second image data (V) is of a second datatype, different from the first datatype; determining (S3) first labeled image data (IL) by labeling the first image data (I) and determining second synthesized labeled image data (ISL) by synthesizing and labeling the second image data (V); training (S4) the neural network based on the received first image data (I), the received second image data (V), the determined first labeled image data (IL) and the determined second labeled synthesized image data (ISL); wherein the first image recognition task and the second image recognition task relate to a same anatomic region where the respective image data is taken from and/or relate to a same pathology to be recogni

Abstract: An apparatus for detecting fraudulent signature inputs is disclosed. The apparatus includes at least a processor and a memory. The memory instructs the processor to receive a plurality of image data from a user. The memory instructs the processor to identify a plurality of signature elements as a function of the plurality of signature inputs. The memory instructs the processor to determine a plurality of signature scores as a function of the plurality of signature elements, wherein the plurality of signature scores comprises a first set of signature scores and a second set of signature scores. The memory instructs the processor to generate an accuracy threshold as a function of the first set of signature scores. The memory instructs the processor to determine one or more fraudulent signature inputs from the plurality of signature inputs as a function of a comparison of signature score to an accuracy threshold.

Abstract: A device implementing a system for machine-learning based gesture recognition includes at least one processor configured to, receive, from a first sensor of the device, first sensor output of a first type, and receive, from a second sensor of the device, second sensor output of a second type that differs from the first type. The at least one processor is further configured to provide the first sensor output and the second sensor output as inputs to a machine learning model, the machine learning model having been trained to output a predicted gesture based on sensor output of the first type and sensor output of the second type. The at least one processor is further configured to determine the predicted gesture based on an output from the machine learning model, and to perform, in response to determining the predicted gesture, a predetermined action on the device.

Abstract: The present disclosure provides an image recognition method and apparatus, an electronic device and a readable storage medium, and relates to the field of artificial intelligence technologies, such as image processing and deep learning technologies. The image recognition method includes: acquiring a to-be-recognized image, and determining a to-be-recognized subject in the to-be-recognized image; extracting a subject feature of the to-be-recognized subject, and obtaining a target feature according to the subject feature; determining a target candidate feature in a plurality of candidate features using the target feature; and taking a class corresponding to the target candidate feature as a recognition result of the to-be-recognized subject. With the present disclosure, different image recognition requirements may be met, and a speed and accuracy of image recognition may be improved.

Abstract: The present disclosure provides an image classification method for maximizing mutual information, device, medium and system, the method including: acquiring a training image; maximizing the mutual information between the training image and a neural network architecture, and automatically determining the network architecture and parameter of the neural network; and processing image data to be classified using the obtained neural network to obtain an image classification result. According to the present disclosure, the network architecture and parameter of the neutral network are automatically designed and determined by maximizing the mutual information based on given image data without burdensome manual design and saving human and computational resource consumption. The present disclosure can automatically design and obtain a neural network-based image classification method in a very short time, and at the same time can achieve higher image classification accuracy.

Abstract: Disclosed in the present disclosure is a method, an apparatus, a system, a device, a storage medium, and a program product for transmitting composite photo data, the method including: responding to a first transmission request triggered by a first user; obtaining target composite photo data directed by the first transmission request, in which the target composite photo data includes at least one piece of first photo data and at least one piece of first video data corresponding to the first photo data; compressing and processing the target composite photo data according to the first photo data to obtain first compressed data having an identical data facade to that of the first photo data; and transmitting the first compressed data to a first terminal position corresponding to the first transmission request.

Abstract: A method and system architecture for designing a compressive sensing matrix for machine learning includes receiving an image associated with a classification task and; generating a sensing matrix. The sensing matrix includes an array of nonzero elements of the image. A prism array of prism elements is in communication with the sensing matrix. A row of values corresponding with an input angle of the prism array is mapped to a respective column corresponding with a detector. Then the detector detects light refracted at an output angle dictated by the physical shape of the prism element. A physical model of the detector is fabricated and generates a compressed representation of the image. A machine learning classification algorithm is applied to the compressed representation of the image and generates an optimized non-invertible final determination of the image.

Abstract: A system comprising: a document holder; an image sensor directed toward the document holder; a polarized filter between the image sensor and the document holder; a filter motor configured to rotate the polarized filter; one or more processors; and at least one memory having stored thereon computer program code that, when executed by the one or more processors, instructs the one or more processors to: control the filter motor to rotate the polarized filter; control the image sensors to capture a plurality of images of a document within the document holder, the plurality of images corresponding to a respective relative orientations of the polarized filter to the document; identify a feature of the document; and output, in response comparing respective visual characteristics of the feature of the document to corresponding expected visual characteristics of the first feature for the document, an indication of a verification of the document.

Abstract: An item classification system for use in an item handling system is disclosed.

Abstract: There are a medical image processing apparatus and a medical image processing method that are capable of maintaining the estimation accuracy of noise intensity even though a system noise ratio rises. A medical image processing apparatus includes a difference image generating unit that divides projection data obtained by applying radiation to an examinee to create a difference image between tomographic images reconstructed for every divided piece of projection data; a local variance computing unit that computes local variance in the difference image; and a noise estimation unit that corrects the local variance using a correction function found in advance to estimate noise intensity of a tomographic image of the examinee. The correction function includes system noise.

Abstract: The size estimation device 5 is provided with an estimation unit 51 and a calculation unit 52. The estimation unit 51 estimates, on the basis of a learning model which learns the feature of a first underwater creature body that faces one direction in an image captured in the water, a first feature point that indicates one end in a one direction side of a second underwater creature body that faces a one direction included in an image captured outside and a second feature point that indicates the other end. The calculation unit 52 calculates information indicating the size of the second underwater creature body on the basis of information including the first feature point and the second feature point.

Abstract: A method includes obtaining, using a stationary sensor of an electronic device, multiple image frames including first and second image frames. The method also includes generating, using multiple previously generated motion vectors, a first motion-distorted image frame using the first image frame and a second motion-distorted image frame using the second image frame. The method further includes adding noise to the motion-distorted image frames to generate first and second noisy motion-distorted image frames. The method also includes performing (i) a first multi-frame processing (MFP) operation to generate a ground truth image using the motion-distorted image frames and (ii) a second MFP operation to generate an input image using the noisy motion-distorted image frames.

Abstract: A system for processing images captured in a retail store and automatically identifying misplaced products is provided. The system may comprise at least one processor configured to receive one or more images captured by one or more image sensors from an environment of a retail store, detect in the one or more images a first product, determine that the first product is not located in the first correct display location, cause an issuance of a user-notification associated with the first product, detect in the one or more images a second product, determine that the second product is not located in the second correct display location, and after determining that the second product is not located in the second correct display location and when the second urgency level is lower than the first urgency level, withhold issuance of a user-notification associated with the second product.

Abstract: In various examples, the present disclosure relates to using temporal filters for automated real-time classification. The technology described herein improves the performance of a multiclass classifier that may be used to classify a temporal sequence of input signals—such as input signals representative of video frames. A performance improvement may be achieved, at least in part, by applying a temporal filter to an output of the multiclass classifier. For example, the temporal filter may leverage classifications associated with preceding input signals to improve the final classification given to a subsequent signal. In some embodiments, the temporal filter may also use data from a confusion matrix to correct for the probable occurrence of certain types of classification errors. The temporal filter may be a linear filter, a nonlinear filter, an adaptive filter, and/or a statistical filter.