Abstract: A series of steps may be performed to automatically extract structured data from an image. First, anchor points may be extracted from the image, representing areas of the image that potentially contain information of interest. The arrangement of anchor points may be used to identify a template. A transform may be generated, to facilitate mapping between particular points in the selected template and corresponding points in the image. The transform may then be used to automatically read visual information from the image and extract structured data from the visual information.

Abstract: Methods, apparatuses, and embodiments related to analyzing the content of digital images. A computer extracts multiple sets of visual features, which can be keypoints, based on an image of a selected object. Each of the multiple sets of visual features is extracted by a different visual feature extractor. The computer further extracts a visual word count vector based on the image of the selected object. An image query is executed based on the extracted visual features and the extracted visual word count vector to identify one or more candidate template objects of which the selected object may be an instance. When multiple candidate template objects are identified, a matching algorithm compares the selected object with the candidate template objects to determine a particular candidate template of which the selected object is an instance.

Abstract: A neural network includes a drop layer configured to drops feature values. A method of computation using the neural network includes extracting feature data from input data using a first portion of a neural network, generating compressed representation data of the extracted feature data by dropping a feature value from the extracted feature data at a drop layer of the neural network based on a drop probability corresponding to the feature value, and indicating an inference result from the compressed representation data using a second portion of the neural network.

Abstract: Disclosed herein are platforms, systems, devices, methods and media for model optimization and data analysis using machine learning. Input data can be processed and analyzed to identify relevant discriminating features, which can be modeled using a plurality of machine learning models. A computing device can be configured with one or more optimized models for categorizing input data.

Abstract: Image transformation tasks such as cropping, text addition etc. are common across industries. Each industry has different business context and demands the image transformations be performed aligned to the business context. This disclosure relates to a system and method for an adaptive image transformation for a given context and maintaining aesthetic sense of the transformed image. Herein, the system is configurable and adaptive to any business context or domain. The system learns the context from available domain samples and creates an automated workflow of context-aware transformation tasks that maintains both the content and aesthetics demands of the context. Further, a saliency map is extracted for the identified RoI to append a text to the RoI based on the extracted saliency map, the calculated similarity metric for various content and aesthetic factors and various preferences of the user.

Abstract: The present disclosure describes imaging systems configured to generate adaptive scanning protocols based on anatomical features and conditions identified during a prenatal scan of an object. Systems may include an ultrasound transducer configured to acquire echo signals responsive to ultrasound pulses transmitted toward a target region. Processors coupled with the transducer can generate an image frame from the echoes and provide the image frame to a first neural network. The first neural network may be configured to identify an anatomical feature in the image frame. An indication of the anatomical feature may be provided a second neural network. The second neural network may then determine an anatomical measurement to be obtained based, in part, on the feature identified. The processors may be further configured to cause an indicator of the anatomical measurement to be obtained to be displayed on a user interface.

Abstract: Methods, systems, and computer readable media for providing personalized saliency models, e.g., for use in mixed reality environments, are disclosed herein, comprising: obtaining, from a server, a first saliency model for the characterization of captured images, wherein the first saliency model represents a global saliency model; capturing a first plurality of images by a first device; obtaining information indicative of a reaction of a first user of the first device to the capture of one or more images of the first plurality images; updating the first saliency model based, at least in part, on the obtained information to form a personalized, second saliency model; and transmitting at least a portion of the second saliency model to the server for inclusion into the global saliency model. In some embodiments, a user's personalized (i.e., updated) saliency model may be used to modify one or more characteristics of at least one subsequently captured image.

Abstract: Various embodiments of the present invention provide methods, apparatuses, systems, computing devices, computing entities, and/or the like for monitoring and/or controlling processing parameters for an industrial process. In various embodiments, a method is provided that comprises: receiving media of a processing region of an industrial process that comprises an object, wherein the media comprises media elements, and each media element comprises a field of view of the object; identifying a set of pixels found in the area of interest; and for each media element: extracting an attribute value from each pixel found in the media element; constructing an attribute profile comprising the attribute value for each pixel; mapping the attribute profile to a mapped profile that comprises at least one property value that correlates to at least one attribute value; and providing the mapped profile to a control system to use in controlling processing parameters of the industrial process.

Abstract: Method and system for classifying and labeling images, which can perform segmentation based on features of each part of images, classify and match the image and the segmented image based on a classification model built by the machine learning method. Meanwhile, each image is assigned with labels and text descriptions. The system also includes a string module assigning the image with a plurality of matching labels and text descriptions that are the most relevant in recent times. Furthermore, the classification model is trained by machine learning method such as an unsupervised learning, a self-supervised learning, or a heuristic algorithms. In addition, a character recognition module is provided to extract characters in the image for comprehensive learning and calculations to facilitate classification and labeling of the image.

Abstract: Systems and methods for evaluating face recognition algorithms in view of image classification features affected by smart makeup are provided. Increasingly, face recognition technology is being used in applications beyond biometric identification for authentication/login purposes. Face recognition technology has been deployed as part of surveillance cameras, which may capture facial images that can be used as evidence of criminal conduct in a court of law. While admissibility of such evidence is subject to the rules of evidence used for any other piece of evidence, reliance on such face recognition technology poses challenges. Until these face recognition algorithms become properly trained and are deployed in a manner that does not result in false positives in the context of policing and judiciary, other solutions are needed. Smart makeup will improve the usage of this technology that has traditionally had a higher false positive identification rate for people of color and women.

Abstract: According to the present invention, proposed is a method for collecting filtered image data, the method comprising the steps of: obtaining, by a computing apparatus, first image data and recording the first image data in an image data pool; obtaining, by the computing apparatus, second image data; performing, by the computing apparatus, an operation in a deep learning model by using the first image data and the second image data as input values, and calculating a first feature vector corresponding to the first image data and a second feature vector corresponding to the second image data; and comparing, by the computing apparatus, a similarity between the first feature vector and the second feature vector, and when the similarity is less than a certain value, recording the second image data in the image data pool.

Abstract: A human behavior recognition method, a device, and a storage medium are provided, which are related to the field of artificial intelligence, specifically to computer vision and deep learning technologies, and applicable to smart city scenarios. The method includes: obtaining attribute information of a target object and N pieces of candidate behavior-related information of a target human from a target image, wherein N is an integer greater than or equal to 1; determining target behavior-related information based on comparison results between the N pieces of candidate behavior-related information and the attribute information of the target object; and determining a behavior recognition result of the target human based on the target behavior-related information.

Abstract: In an image processing-based system for measuring moisture and light transmittance and a method for measuring moisture and light transmittance using the same, the image processing-based system includes a coating paper, a photographing box, a microscope and an RGB decision program. The coating paper is coated with coating conditions substantially same as the coating conditions for a metal material. The photographing box is configured to photograph the coating paper, to obtain a first image. The microscope is configured to photograph the coating paper, to obtain a second image. The RGB decision program is configured to obtaining an RGB value of each of the first and second images. Moisture transmittance is evaluated using the RGB value of the first image, and light transmittance is evaluated using the RGB value of the second image.

Abstract: Described herein are systems and methods for determining if a vehicle is parked. In one example, a system includes a processor, a sensor system, and a memory. Both the sensor system and the memory are in communication with the processor. The memory includes a parking determination module having instructions that, when executed by the processor, cause the processor to determine, using a random forest model, when the vehicle is parked based on vehicle estimated features, vehicle learned features, and vehicle taillight features of the vehicle that are based on sensor data from the sensor system.

Abstract: A system is provided for identifying and quantizing a change in a scene having objects based on comparison of features representing a target object derived from a first image and a second image that includes the target object. The system segments, based on a scene model of the scene, a second image that includes the target object to identify a second area representing the target object within the second image. The system extracts, from the second image, second features for the target object based on the second area. The system determines a second effectiveness of the second features indicating effectiveness of the second features in representing the target object. The system detects a change based on a difference between first features and the second features factoring in a first effectiveness and the second effectiveness.

Abstract: A retrieval apparatus includes a first retrieval unit, a second retrieval unit, and an integration unit that calculates an integrated similarity by integrating a first similarity calculated by the first retrieval unit and a second similarity calculated by the second retrieval unit. For similarities between a basic image, an image similar to the basic image, and an image dissimilar to the basic image of the reference images, at least the feature extraction of the first retrieval unit is learned such that a margin based on a second similarity between the basic image and the similar image and a second similarity between the basic image and the dissimilar image increases as the second similarity between the basic image and the dissimilar image increases relative to the second similarity between the basic image and the similar image.

Abstract: The welding operation monitoring system includes an image capturing device that is disposed on a side opposite to a side to which the plasma flow is supplied out of a pipe inside and a pipe outside of the strip-shaped steel sheet formed in a tubular shape, and is configured to capture a color image including a plasma flow over the V-shaped region; and a welding operation monitoring device that is configured to generate a specific color component image obtained by extracting a specific color component from the color image, and specifies a V-shaped display region, which is a region corresponding to the V-shaped region within the color image, on the basis of the V-shaped region shown in the specific color component image, thereby analyzing a state of the welding operation.

Abstract: This disclosure generally relates to method and system for asset inspection using unmanned aerial vehicles. Regular inspection of assets in any environment is time consuming and involves physical risk and injury to the property. The method includes receiving a plurality of input images of an asset using unmanned aerial vehicle and generates a 3D point cloud based on overlapped images identified from the plurality of input images of the asset being inspected. Further, anomalies are detected and defects are observed in the one or more objects of interests based on the change observed in each 2D image frame from the plurality 2D image frames in a specific view with an identical objects of interest of the 2D image frame. A pyramidal view map is created for the plurality 2D image frames by mapping each pixel of the plurality 2D image frames with the plurality of input images.

Abstract: A method to determine activity in a sequence of successively acquired images of a scene, comprises: acquiring the sequence of images; for each image in the sequence of images, forming a feature block of features extracted from the image and determining image specific information including a weighting for the image; normalizing the determined weightings to form a normalized weighting for each image in the sequence of images; for each image in the sequence of images, combining the associated normalized weighting and associated feature block to form a weighted feature block; passing a combination of the weighted feature blocks through a predictive module to determine an activity in the sequence of images; and outputting a result comprising the determined activity in the sequence of images.

Abstract: Disclosed are methods, systems, and computer-readable medium for obstacle detection and database management using multi-source weightages. For instance, the method may include: responsive to a detection of an object, determining whether the object is a known object or an unknown object based on object features and known objects indicated by vehicle state information and an obstacle database; responsive to a determination that the object is an unknown object, updating the obstacle database with unknown object information and transmitting the unknown object information to an off-board service; receiving a response from the off-board service, the response including a weightage assigned to the unknown object; updating the obstacle database with the weightage assigned to the unknown object; and performing at least one action based on the weightage of the unknown object.