Abstract: System, methods, and other embodiments described herein relate to an improved approach to training a depth model for monocular depth estimation by warping depth features prior to decoding. In one embodiment, a method includes encoding, using an encoder of a depth model, a source image into depth features of a scene depicted by the source image. The method includes warping the depth features into warped features of a target frame of a target image associated with the source image. The method includes decoding, using a decoder of the depth model, the warped features into a depth map. The method includes training the depth model according to a loss derived from the depth map.

Abstract: A single credential capture event enables a plurality of persons to transit a physical access control portal. A cohort is determined when each of a plurality of persons is found within an identity indicia store. When a credential for any one of the cohort is validated, a physical access control actuator is enabled. A video stream of the plurality of persons is transformed to a plurality of best face images. When each one of the best face images is correlated with an entry in the identity indicia store above a threshold, a potential validation window is opened for the cohort. When a credential event for any of the plurality of persons occurs within the validation window, admittance is enabled through the physical access control portal. A cohort is a contiguous sequence of persons which begins with a first correlated best face and is terminated before the first uncorrelated best face.

Abstract: This disclosure provides a method for imaging lymph nodes and lymphatic vessels without a contrast agent. The method includes providing, using an optical source, an infrared illumination to a region of a subject having at least one lymphatic component, detecting a reflected portion of the infrared illumination directly reflected from the region using a sensor positioned thereabout, and generating at least one image indicative of the at least one lymphatic component in the subject using the reflected portion of the infrared illumination.

Abstract: A vision-based target tracking method, a system, an equipment, and a storage medium. The method includes: acquiring face positions and body positions in a current image frame; acquiring a current position of a time-series tracking target by a time-series tracking algorithm; determining, when a face matching the face of a to-be-tracked target does not exist outside a region corresponding to the current position of the time-series tracking target, whether the time-series tracking target is in an intersecting state; determining whether a body matching the body of the to-be-tracked target in other persons intersecting the times-series tracking target when in the intersecting state, or using the current position of the time-series tracking target as a current target position if not in the intersecting state; and taking the current target position as the position of the to-be-tracked target in the previous frame, and returning to the above steps.

Abstract: A method for determining a salient object in an image based on superpixel analysis consists of initializing superpixels using SLIC algorithm and merging of adjacent superpixels that have similar color distribution. Then, calculate the spatial and color distribution correlation between the superpixels after being merged. Combined with the statistics of the occupancy rate, the distance to the original image center, and the global contrast of the superpixels, calculate the saliency evaluation vector for each superpixel. Finally, interpolate the saliency for each pixel in the superpixel.

Abstract: A method of counting pigs or poultry includes mounting a camera at an alleyway through which the livestock pass with a width such that the livestock can pass at least two abreast and are free to turn to move in an opposite direction against the direction of travel before all livestock of the group complete their movement though the alleyway. An analysis of the images is carried out which takes into account errors caused by livestock which turn in the portion of the alleyway to move in the opposite direction temporarily before continuing to complete the travel through the portion of the single alleyway to the remote location and takes into account errors caused animals being side by side when passing through the portion of the single alleyway.

Abstract: A method for training a depth estimation model is provided. The method includes obtaining a first left image and a first right image. A disparity map is obtained by inputting the first left image into a depth estimation model. A second right image is obtained by adding the first left image to the disparity map. The first left image is converted into a third right image. A mask image is obtained by performing a binarization processing on a pixel value of each of pixel points of the third right image. Once a loss value of the depth estimation model is obtained by calculating a mean square error of pixel values of all corresponding pixel points of the first right image, the second right image, and the mask image, a depth estimation model is iteratively trained according to the loss value.

Abstract: A method in a mobile computing device includes: controlling a camera to capture an image; tracking a pose of the mobile computing device, corresponding to the image, in a coordinate system; detecting an item in the image; determining a location of the detected item in the coordinate system, based on the tracked pose; obtaining an item identifier corresponding to the detected item, based on the location of the detected item in the coordinate system; generating a training data sample including (i) a payload based on the detected item, and (ii) a label including the obtained item identifier; and storing the training data sample.

Abstract: Embodiments of the present invention relates to a stereo matching method, an image processing chip and a mobile vehicle. The stereo matching method includes: calculating aggregate cost values between all reference pixels and a target pixel in a preset search region, the reference pixel being a pixel in a reference image, and the target pixel being a pixel in a target image; determining a texture property of the reference image in the search region according to a distribution of the aggregate cost values; and using a method for calculating a disparity value corresponding to the texture property, to obtain a disparity value between the reference image and the target image at a position of the target pixel.

Abstract: A method of displaying a digital object on a device is disclosed. A set of unique identifiers and associated is downloaded for a set of digital objects within a 3D model. A position and an orientation of a virtual camera is determined relative to the positions of the set of digital objects. A visibility score for each digital object of the set of digital objects is computed based on a position of the digital object relative to the position and the orientation of the virtual camera. A polygon mesh associated with a digital object with a highest visibility score is downloaded based on the polygon mesh being absent. The downloading uses a unique identifier of the set of unique identifiers associated with the polygon mesh. The polygon mesh of the digital object with the highest score is rendered and displayed on a display device.

Abstract: An image processing device includes: an obtaining unit configured to obtain an image including at least one body hair; and an analyzing unit configured to analyze the image to output information indicating an orientation of the at least one body hair included in the image.

Abstract: Embodiments described herein provide systems, methods, and devices for pre-training a multimodal encoder-decoder (MED) model for vision-language tasks. A method may include encoding, by an image encoder of the MED, an image into an image representation; encoding, by a text encoder of the MED, a text into a text representation; generating, by an image-grounded text encoder of the MED, a multimodal representation based on the image representation and the text; generating, by an image-grounded text decoder of the MED, a predicted text based on the image representation and the text; generating, through an image-text matching (ITM) head, a binary classification indicating whether the image and the text are a match; computing a first loss, ITM loss, and third loss based on the image representation, text representation, binary classification, predicted text and text; jointly updating the MED based on the first loss, the second loss and the third loss.

Abstract: Techniques for color change of Information elements (IEs) are described. In an example, a color of the IE and a color of a region of a virtual environment surrounding the IE may be compared. Based on the comparison, it may be determined whether the IE is distinguishable from the region of the virtual environment. Further, based on the determination, the color of the IE may be changed to facilitate distinguishability of the IE.

Abstract: A method and system for remotely detecting a heart-related condition is presented. The method includes receiving a facial image of a patient, wherein the facial image is retrieved from a data store and captured from the patient; extracting at least one facial feature from the facial image of the patient, wherein the at least one facial feature is represented as numerical descriptors of facial attributes, wherein the numerical descriptors include surface features of the at least one facial feature; and classifying the extracted at least one facial feature using a classifier, wherein the classifier maps the extracted at least one facial feature to at least one score indicative of a stage of the heart-related condition.

Abstract: Provided are a memory and a processor connected to the memory to execute a program stored in the memory. The processor classifies each of pixels included in an image into any one region of a region of achromatic color and a region of chromatic color inside a color space, uses the pixels classified into the region of achromatic color to extract a first predetermined number of representative colors, uses the pixels classified into the region of chromatic color to extract a second predetermined number of representative colors, and replaces each of the pixels included in the image with a pixel of a representative color having a smallest difference in pixel value from the pixels included in the image, among the first predetermined number of representative colors and the second predetermined number of representative colors, to generate a color-reduced image of the image.

Abstract: The disclosure herein relates to systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking. In some embodiments, the systems, devices, and methods described herein are configured to analyze non-invasive medical images of a subject to automatically and/or dynamically identify one or more features, such as plaque and vessels, and/or derive one or more quantified plaque parameters, such as radiodensity, radiodensity composition, volume, radiodensity heterogeneity, geometry, location, perform computational fluid dynamics analysis, facilitate assessment of risk of heart disease and coronary artery disease, enhance drug development, determine a CAD risk factor goal, provide atherosclerosis and vascular morphology characterization, and determine indication of myocardial risk, and/or the like.

Abstract: A Smart Optical Character Recognition (SOCR) Trainer comprises software developed for automating Quality Control (QC) using unsupervised machine-learning techniques to analyze, classify, and optimize textual data extracted from an image or PDF document. SOCR Trainer serves as a ‘data treatment’ utility service that can be embedded into data processing workflows (e.g., data pipelines, ETL processes, data versioning repositories, etc.). SOCR Trainer performs a series of automated tests on the quality of images and their respective extracted textual data to determine if the extraction is trustworthy. If deficiencies are detected, SOCR Trainer will analyze certain parameters of the document, perform conditional optimizations, re-perform text extraction, and repeat QA testing until the output meets desired specifications. SOCR Trainer will produce audit files recording the provenance and differences between original documents and enhanced optimized document text.

Abstract: A system and method for a procedure that can be performed on an appropriate subject. Procedures can include assembling any appropriate work piece or installing members into a work piece, such as an airframe, autoframe, etc. Regardless of the subject, generally the procedure can have a selected result that is efficacious. The efficacious result may be the desired placement of a device. The system and method can be used in confirming an efficacious result.

Abstract: One aspect provides a method for object detection including detecting, using an electronic processor, a plurality of candidate objects in a video using a convolutional neural network detection process and a background subtraction detection process and identifying, using the electronic processor, a candidate object from the plurality of candidate objects. The candidate object detected by the background subtraction detection process in a location of the video with no candidate objects detected by the convolutional neural network detection process. The method also includes determining, using the electronic processor, a background subtraction confidence level of the candidate object and categorizing, using the electronic processor, the candidate object as a detected object in the video in response to the background subtraction confidence level satisfying a background subtraction confidence threshold.

Abstract: A worksite risk analysis and documentation system and method defines, captures, categorizing, and documents, while analyzing functional physical demands of a plurality of jobs, tasks, body movements to provide musculoskeletal and health and safety risks at various worksites. The system uses the use of odds ratios in the determination of risk for musculoskeletal, ergonomic and EHS&S risk factors. A built in QR tracking system for Covid-19 is provided.