Abstract: Electronic devices, methods, and non-transitory program storage devices for leveraging machine learning to perform improved image fusion and/or noise reduction are disclosed. When a capture request is received, a neural network may be used to perform fusion and denoising operations on a first set of captured input images. According to some embodiments, the neural network's architecture comprises: a first plurality of network layers configured to compute optical flow information between the first set of input images; a second plurality of network layers configured to perform, at least in part, the fusion and denoising operations on the first set of input images; and a third plurality of skip connections between layers of the second plurality of network layers, wherein parameters for each skip connection of the third plurality of skip connections are warped based on at least part of the optical flow information computed by the first plurality of network layers.

Abstract: A method for estimating a stimulated emission depletion microscopy (STED) resolution includes generating a first frame representing a reference image from a field-of-view, the reference image having a predetermined reference resolution, and generating at least one second frame representing a STED image from the same field-of-view, the STED image having the STED resolution to be estimated. The at least one second frame is blurred by applying a convolution kernel with at least one fit parameter to the second frame. An optimal value of the at least one fit parameter of the convolution kernel is determined for which a difference between the first frame and the blurred at least one second frame is minimized. The STED resolution is estimated based on the optimal value of the at least one fit parameter and the predetermined reference resolution.

Abstract: An optical fingerprint sensor with spoof detection includes a plurality of lenses, an image sensor including a pixel array that includes a plurality of first photodiodes and a plurality of second photodiodes, and at least one apertured baffle-layer having a plurality of aperture stops, wherein each second photodiode is configured to detect light having passed through a lens and at least one aperture stop not aligned with the lens along an optical axis. A method for detecting spoof fingerprints detected using an optical fingerprint sensor includes detecting large-angle light incident on a plurality of anti-spoof photodiodes, wherein the plurality of anti-spoof photodiodes is interleaved with a plurality of imaging photodiodes, determining an angular distribution of light based at least in part one the large-angle light, and detecting spoof fingerprints based at least in part on the angular distribution of light.

Abstract: Disclosed is a method and apparatus for testing a liveness, where the liveness test method includes receiving a color image and a photodiode (PD) image of an object from an image sensor comprising a pixel formed of a plurality of PDs, preprocessing the color image and the PD image, and determining a liveness of the object by inputting a result of preprocessing the color image and a result of preprocessing the PD image into a neural network.

Abstract: A variation of a method for estimating a quantity of a blood component in a fluid canister includes: within an image of a canister, identifying a reference marker on the canister; selecting an area of the image based on the reference marker; correlating a portion of the selected area with a fluid level within the canister; estimating a volume of fluid within the canister based on the fluid level; extracting a feature from the selected area; correlating the extracted featured with a concentration of a blood component within the canister; and estimating a quantity of the blood component within the canister based on the estimated volume and the concentration of the blood component within the canister.

Abstract: A processor-implemented method implementing a convolution neural network includes: determining a plurality of differential groups by grouping a plurality of raw windows of an input feature map into the plurality of differential groups; determining differential windows by performing, for each respective differential group of the differential groups, a differential operation between the raw windows of the respective differential group; determining a reference element of an output feature map corresponding to a reference raw window among the raw windows by performing a convolution operation between a kernel and the reference raw window; and determining remaining elements of the output feature map by performing a reference element summation operation based on the reference element and each of a plurality of convolution operation results determined by performing respective convolution operations between the kernel and each of the differential windows.

Abstract: A method, apparatus and computer program utilize digital signatures to accurately and efficiently identify changes to objects of interest within captured images. In the context of a method, a digital signature is obtained for an object of interest at a particular location. The method also determines whether a difference between the digital signature and a previously determined digital signature associated with the location satisfies a predefined criterion. In accordance with determining that the difference fails to satisfy the predefined criterion, the method causes information regarding the object of interest at the particular location to be collected, such as to permit a map that references the object of interest or other information associated with the object of interest to be updated.

Abstract: In an example implementation according to aspects of the present disclosure, a system, method, and storage medium for processing digital image file metadata characteristics. The system includes a processor and memory. The processor executes machine readable instructions to receive a set of digital image files from a first client application. The processor extracts a set of metadata from the set of digital image files. The processor identifies a set of common characteristics within the set of metadata. The processor groups the set of digital image files based on the set of common characteristics. The processor evaluates an association between the set of common characteristics and a second user. Responsive to the evaluation, the processor sends the grouped set of digital image files to a second client application.

Abstract: Biometric input, such as an image of a hand, may be processed to determine embedding vector data that may be used to identify users. Accuracy of the identification is improved by using high resolution inputs to a deep convolutional neural network (DCNN) that is trained to generate the embedding vector data that is representative of features in the input. Training data sets are expensive to develop and thus may be relatively small. During training of the DCNN, confidence loss values corresponding to the entire input as well as particular patches or portions of the input are determined. These patch-wise confidence loss values mitigate potential overfitting during training of the DCNN and improve overall performance of the trained DCNN to determine embedding vector data suitable for identification.

Abstract: Approximate modeling of next combined result for stopping text-field recognition in a video stream. In an embodiment, text-recognition results are generated from frames in a video stream and combined into an accumulated text-recognition result. A distance between the accumulated text-recognition result and a next accumulated text-recognition result is estimated based on an approximate model of the next accumulated text-recognition result, and a determination is made of whether or not to stop processing based on this estimated distance. After processing is stopped, the final accumulated text-recognition result may be output.

Abstract: This disclosure relates to a method and system for extracting information from images of one or more templatized documents. A knowledge graph with a fixed schema based on background knowledge is used to capture spatial and semantic relationships of entities present in scanned document and an adaptive lattice-based approach based on formal concepts analysis (FCA) is used to determine a similarity metric that utilizes both spatial and semantic information to determine if the structure of the scanned document image adheres to any of the known document templates. If a known document template whose structure is closely matching the structure of the scanned document is detected, then an inductive rule learning based approach is used to learn symbolic rules to extract information present in scanned document image and if a new document template is detected, then future scanned document images belonging to new document template are automatically processed using the learnt rules.

Abstract: This disclosure describes techniques for identifying users that are enrolled for use of a user-recognition system and updating identification data of these users over time. To enroll in the user-recognition system, the user may initially scan his or her palm. The resulting image data may later be used when the user requests to be identified by the system by again scanning his or her palm. However, because the characteristics of user palms may change over the time, the user-recognition system may periodically perform processes for updating the identification data stored in association with the user in order to maintain or increase an accuracy of the user-recognition system.

Abstract: Video containers are expanded to encapsulate auxiliary data and metadata linking video content with relevant portions of auxiliary data for download during playback of particular video content. Auxiliary data is parsed into areas of interest and associated with the video content including the same areas of interest. During playback a viewer is provided access to relevant portions of the auxiliary data according to the encapsulated metadata describing the relevant portions.

Abstract: Various methods and systems are provided for individually analyzing a plurality of subregions within an ultrasound image for acoustic shadowing. In one embodiment, a method includes acquiring ultrasound data along a plurality of receive lines, generating an ultrasound image based on the ultrasound data, dividing the ultrasound image into a plurality of subregions, and individually analyzing each of the plurality of subregions for acoustic shadowing. The method includes detecting acoustic shadowing in one or more of the plurality of subregions, displaying the ultrasound image, and graphically indicating the one or more of the plurality of subregions in which the acoustic shadowing was detected on the ultrasound image while the ultrasound image is displayed on a display device.

Abstract: A method for obtaining image data of a subject, including; a first scanning step including a plurality of steps W, each step W being a step of determining one place existing in a one-dimensional, two-dimensional, or three-dimensional first space; and a second scanning step of scanning insides of second spaces including at least one of the places, wherein the second scanning step includes a step X of randomly determining a location of an observation point and a step Y of obtaining a piece of image data for the observation point, and, at a time point of end of scanning an inside of one of the second spaces, the second space has a first region including 50% of observation points and a second region existing outside the first region and including remaining 50% of the observation points, the second region being larger than the first region by at least 15%.

Abstract: The present disclosure provides a method and device for correcting a scanned image, and an image scanning system, and relates to the field of image scanning. The method includes obtaining a scanned image of a scanned object, detecting one or more reference objects from the scanned image, determining a deformation parameter of each reference object of the one or more reference objects based on preset a standard parameter of the each reference object, and correcting the scanned image based on the deformation parameters of the one or more reference objects.

Abstract: Executable code causes a processor to segment a first frame to determine a background portion of the first frame, and segment a second frame to determine a background portion of the second frame. The executable code causes the processor to compare the background portion of the first frame to the background portion of the second frame to determine a difference between the first frame and the second frame. The executable code also causes the processor to alter the background portion of the second frame responsive to the difference. The executable code causes the processor to display the altered second frame on a display.

Abstract: The present disclosure provides a method for detecting the focus of attention. The method includes: obtaining the face of a person in the first image, as well as the result of facial recognition; determining whether the distance between the person and the target is within an effective attention range; determining whether the face is frontal; determining whether the effective attention period is not shorter than a period threshold; detecting the focus of attention for the person to the target.

Abstract: A document processing apparatus including a processor configured to receive data including plural document sets each made up of predetermined plural pages; classify the data into sets; and in a case where the data includes an unclassified page, display a candidate for a set to which the unclassified page belongs.

Abstract: In an example embodiment, machine learning is used to intelligently compress documents to reduce the overall footprint of storing large amounts of files for an organization. Specifically, a document is split into parts, with each part representing a grouping of text or an image. Optical character recognition is performed to identify the text in images. Machine learning techniques are then applied to a part of a document in order to determine how relevant the document is for the organization. The parts that are deemed to be not relevant may then be reduced in size, either by omitting them completely or by summarizing them. This allows for the compression to be tailored specifically to the organization, resulting in the ability to compress or eliminate parts of documents that other organizations might have found relevant (and thus would not have been compressed or eliminated through traditional means).