Abstract: Techniques to train a model with machine learning and use the trained model to select a bounding box that represents an object are described. For example, a system may implement various techniques to generate multiple bounding boxes for an object in an environment. Each bounding box may be slightly different based on the technique and data used. To select a bounding box that most closely represents an object (or is best used for tracking the object), a model may be trained. The model may be trained by processing sensor data that has been annotated with bounding boxes that represent ground truth bounding boxes. The model may be implemented to select a most appropriate bounding box for a situation (e.g., a given velocity, acceleration, distance, location, etc.). The selected bounding box may be used to track an object, generate a trajectory, or otherwise control a vehicle.

Abstract: Embodiments of the present application provide a character recognition method and device. The method includes obtaining a target image to be analyzed which contains a character (S101); inputting the target image into a pre-trained deep neural network to determine a feature map corresponding to a character region of the target image (S102); and performing character recognition on the feature map corresponding to the character region by the deep neural network to obtain the character contained in the target image (S103). The deep neural network is obtained by training with sample images, a result of labeling character regions in the sample images, and characters contained in the sample images. The method can improve the accuracy of character recognition.

Abstract: The present invention provides a method, system and image to enhance the image contrast of a digital image device while simultaneously compensating for image intensity inhomogeneity, regardless of the source. The present invention corrects intensity inhomogeneities producing a more uniform image appearance. Also, the image is enhanced through increased contrast, e.g., tissue contrast in a medical image. The method makes no assumptions as to the source of the inhomogeneities, e.g., physical device characteristics or positioning of the object being imaged. In the method, the error between the histogram of the spatially-weighted original image and a specified histogram is minimized. The specified histogram may be selected to increase contrast generally or particularly for accentuation, e.g., on localized regions of interest. The weighting is preferably achieved by two-dimensional interpolation of a sparse grid of control points overlaying the image.

Abstract: The invention is a method for characterizing an object (20), comprising the following steps: a) placing the object between a radiation source (10) and a radiation detector (30); b) irradiating the object with the radiation source and detecting radiation transmitted by the object (14) using the radiation detector, the radiation detector defining a plurality of pixels; c) for each pixel (30i), forming an energy spectrum (Si) of the detected radiation, each spectrum comprising at least two distinct energy bands; d) from each spectrum formed in c), estimating, in each pixel, at least two equivalent thicknesses ({circumflex over (L)}i,1 . . . {circumflex over (L)}i,M,{circumflex over (L)}?i,1 . . . {circumflex over (L)}?i,M) respectively associated with at least two basic materials (mat1 . . . matM,mat?1 . . .

Abstract: A method for monitoring a surface of an object via at least one optical camera including the steps: (i) image capture of a predetermined surface segment of the object by a camera at a first point in time, (ii) inspection of the predetermined surface segment of the object such that the surface of the object is acted on, and (iii) image capture of the predetermined surface segment of the object by a camera at a second point in time, which takes place after the first point in time, wherein the image data of the predetermined surface segment of the images of step (i) and step (iii) is sent to an evaluation means and then evaluated by this evaluation means and compared to each other, wherein, if a surface flaw detected in the image capture of step (iii), both in terms of its shape or its position on the surface of the object, coincides with a surface flaw detected in the image capture of step (i), then this surface flaw is classified as an actual surface flaw of the object.

Abstract: The disclosure relates to devices, systems and methods for image registration and annotation. The devices include computer software products for aligning whole slide digital images on a common grid and transferring annotations from one aligned image to another aligned image on the basis of matching tissue structure. The systems include computer-implemented systems such as work stations and networked computers for accomplishing the tissue-structure based image registration and cross-image annotation. The methods include processes for aligning digital images corresponding to adjacent tissue sections on a common grid based on tissue structure, and transferring annotations from one of the adjacent tissue images to another of the adjacent tissue images.

Abstract: In one or more embodiments, a system for calibration between a camera and a ranging sensor comprises a ranging sensor to obtain ranging measurements for a target located at N number of locations with an emitter at M number of rotation positions. The system further comprises a camera to image the target to generate imaging measurements corresponding to the ranging measurements.

Abstract: Systems and methods for performing OCR of a series of images depicting text symbols. An example method comprises performing OCR a series of images to produce a current symbol sequence and corresponding symbol sequence quadrangle; associating the current symbol sequence with a previous symbol sequence for a previously received image; identifying a median string; determining a median symbol sequence quadrangle; and displaying, using the median symbol sequence quadrangle, a resulting OCR text representing at least a portion of the original document.

Abstract: The present technology relates to an image processing apparatus and method that are capable of enhancing encoding efficiency while suppressing a decrease in the efficiency of encoding processing. The image processing apparatus includes an encoding mode setter that sets, in units of coding units having a hierarchical structure, whether a non-compression mode is to be selected as an encoding mode for encoding image data, the non-compression mode being an encoding mode in which the image data is output as encoded data, and an encoder that encodes the image data in units of the coding units in accordance with a mode set by the encoding mode setter. The present disclosure can be applied to, for example, an image processing apparatus.

Abstract: A system includes one or more server hardware computing devices includes one or more processors and memory storing instructions that, when executed by the one or more processors, cause the system to receive a request to verify an identity of a user attempting to complete an examination provided by an online proctored examination platform to a candidate user authorized to complete the examination and repeatedly, at a predefined interval during the examination, determine, based at least in part on user data collected by the online proctored examination platform during the examination, the user data identifying the user at a time that the interval elapses, whether the user is the candidate user, and, responsive to a determination that the user is not the candidate user, perform one or more of a plurality of actions associated with a failed identity verification, the plurality of actions including stopping the examination.

Abstract: Embodiments of the present invention provide a method for processing a radiotherapy CT positioning image, comprising: defining a pixel having a CT value greater than or equal to a first threshold in an original CT image as a human body pixel; counting a number of human body pixels of each pixel row in the original CT image in an order from top to bottom; and determining a boundary of the human body according to the counting result.

Abstract: A method of generating log data that includes a plurality of records, in which information detected through image recognition processing on a plurality of image frames is recorded. The method includes, detecting objects in each image frame through the image recognition processing, generating a record including identification information of each image frame and the aggregated number of objects detected in the image frame for each image frame, and adding metadata relating to the detected objects to the corresponding record when the aggregated number of objects is equal to or greater than one. When the aggregated number of objects is zero, the metadata related to the detected objects is not added to the record.

Abstract: Methods, computer program products, and systems are presented. The methods include, for instance: obtaining a document image, wherein the document image includes a plurality of objects; identifying a plurality of macroblocks within the document image; performing microblock processing within macroblocks of the plurality of macroblocks, wherein the microblock processing includes examining content of microblocks within a macroblock for extraction of key-value pairs, the examining content including performing an ontological analysis of microblocks, wherein the microblock processing includes associating confidence levels to the extracted key-value pairs; and outputting metadata based on the performing microblock processing within macroblocks of the plurality of macroblocks.

Abstract: A depth camera assembly (DCA) determines depth information within a local area. The DCA may selectively process a subset of data captured by an imaging sensor and obtained from the imaging sensor, such as pixels corresponding to a region of interest, for depth information. Alternatively, the DCA may limit retrieval of data from the imaging sensor to pixels corresponding to the region of interest from the imaging sensor for processing for depth information. The depth processing may include a semi-global match (SGM) algorithm, and the DCA adjusts a number of neighboring pixels used for determining depth information for a specific pixel based on one or more criteria. In some embodiments, the DCA performs the depth processing by analyzing images from different image sensors using left to right and right to left correspondence checks that are performed in parallel.

Abstract: A skeleton posture determining method and apparatus, and a computer readable storage medium belong to the field of human body posture recognition, where the method includes obtaining a first skeleton posture including location information of a plurality of joints, determining a description parameter of at least one target joint of the joints, obtaining, through screening, a correct joint from the at least one target joint based on the description parameter, and performing fitting based on location information of the correct joint to obtain a second skeleton posture, where the second skeleton posture includes location information of at least one joint, and the at least one target joint is in a one-to-one correspondence with the at least one joint. Hence, the method, the apparatus, and the computer readable storage medium can improve recognition accuracy of a human body posture.

Abstract: A distributed systems and methods for generating composite media including receiving a media context that defines media that is to be generated, the media context including: a definition of a sequence of media segment specifications and, an identification of a set of remote devices. For each media segment specification, a reference segment may be generated and transmitted to at least one remote device. A media segment may be received from each of the remote device, the media segment having been recorded by a camera. Verified media sequences may replace the corresponding reference segment. The media segments may be aggregated and an updated sequence of media segments may be defined. An instance of the media context that includes a subset of the updated sequence of media segments may then be generated.

Abstract: A best optical inspection mode to detect defects can be determined when no defect examples or only a limited number of defect examples are available. A signal for a defect of interest at the plurality of sites and for the plurality of modes can be determined using electromagnetic simulation. A ratio of the signal for the defect of interest to the noise at each combination of the plurality of sites and the plurality of modes can be determined. A mode with optimized signal-to-noise characteristics can be determined based on the ratios.

Abstract: Techniques are described that facilitate automatically distinguishing between different expressions of a same or similar emotion. In one embodiment, a computer-implemented is provided that comprises partitioning, by a device operatively coupled to a processor, a data set comprising facial expression data into different clusters of the facial expression data based on one or more distinguishing features respectively associated with the different clusters, wherein the facial expression data reflects facial expressions respectively expressed by people. The computer-implemented method can further comprise performing, by the device, a multi-task learning process to determine a final number of the different clusters for the data set using a multi-task learning process that is dependent on an output of an emotion classification model that classifies emotion types respectively associated with the facial expressions.

Abstract: An information processing apparatus includes an acquisition unit that acquires process information associated with a reference image and regarding an input process performed by referencing the reference image, and a process decision unit that decides a content of a next input process to perform by referencing the reference image, on a basis of the process information acquired by the acquisition unit.

Abstract: A method of performing an iterative bundle adjustment for an imaging device is described. The method comprising implementing a plurality of functions in performing a bundle adjustment. Predetermined functions of the plurality of functions may be started using a processor for a second iteration in parallel with a first iteration of the plurality of functions. A result of the predetermined functions started during the first iteration may be used in a second iteration. An output of the bundle adjustment may then be generated for successive iterations.