Abstract: Categorizing images may include training a first neural network to cluster a plurality of images to obtain a first image embedding space, wherein a vector representation is determined for each of the plurality of images based on the training, determining a vector norm value corresponding to each of the plurality of images based on the vector representation for each of the plurality of images, and identifying a first subset of the images for which a corresponding vector norm value satisfies a predetermined vector norm quality threshold. Then, a second neural network may be trained using the first subset of images to obtain a second image embedding space, and the second image embedding space may be used to categorize additional images.

Abstract: A repetitive structure extraction device includes an image feature extraction unit which extracts an image feature for each of a plurality of images which are captured at one or a plurality of locations and which are given different capture times, a temporal feature extraction unit which extracts, for each of the plurality of images, a temporal feature according to a predetermined period from a capture time given to the image, and a repetitive structure extraction unit which learns, on the basis of the image feature extracted for each of the plurality of images by the image feature extraction unit and the temporal feature extracted for each of the plurality of images by the temporal feature extraction unit, a repetitive structure which is used to perform interconversion between the temporal feature and a component of the image feature and which is provided according to a correlation of periodic change between the component of the image feature and the temporal feature.

Abstract: A system and method for aligning image data from a vehicle camera is provided.

Abstract: A Position and Orientation Measurement Engine (POME) is a mobile camera system that can be used for accurate indoor measurement (e.g., at a construction site). The POME uses a plurality of cameras to acquire images of a plurality of targets. If locations of the plurality of targets are precisely known, images of the targets can be used to determine a position of the POME in relation to the plurality of targets. However, to precisely determine locations of the plurality of targets can be time consuming and/or use expensive equipment. This disclosure discusses how to use the POME itself to determine locations of the plurality of targets.

Abstract: The present disclosure relates to a method, storage medium, and system for analyzing an image sequence of a periodic physiological activity. In one implementation, the method includes receiving the image sequence acquired by an imaging device, the image sequence having a plurality of frames, and identifying a feature point in a first frame. The method further includes determining motion vectors for the feature point in the frames of the image sequence. Each motion vector for the feature point is determined based on respective locations of corresponding feature points in frames adjacent to the first frame. The method also includes determining a motion magnitude profile based on the determined motion vectors and determining a phase of each frame in the image sequence based on the motion magnitude profile.

Abstract: If a lesion included in a specification target image is a texture lesion, a probability image calculation unit calculates a probability value indicating a probability that each of a plurality of pixels of the specification target image is included in a lesion area. An output unit calculates, as a candidate area, an area including pixels whose probability values are equal to or larger than a first threshold in a probability image obtained from the probability image calculation unit and, as a modification area, an area including pixels whose probability values are within a certain probability range including the first threshold. An input unit detects an input from a user on a pixel in the modification area. A lesion area specification unit specifies a lesion area on the basis of the probability image, the candidate area, the modification area, and user operation information.

Abstract: An image processing device according to one embodiment includes a processor. The processor executes a step of acquiring an input image, a step of calculating a feature residual by processing the input image in a convolutional layer, a step of performing at least one convolution on the input image, a step of generating an output feature by applying the feature residual to the convolved input image, and a step of generating an image residual based on the output feature. The image residual is applied to the input image, and thereby a high-resolution image with higher resolution than the input image is generated.

Abstract: A method for predicting a future dental situation for a patient. Acquiring historical data including at least: previous time point and context parameter values. Acquiring at a time point all the data related to the current dental situation, including at least: the current time point, context parameter values at the current time point, statistical analysis of the historical data and the current data, so as to predict, at a future time point, at least one future dental situation for the current patient. Depending on the future dental situation, (re)evaluation of the benefit of an orthodontic treatment. Steps may include creating a three-dimensional digital reference model, acquiring at least one two-dimensional image, analysing each updated image and creation, searching, for each updated image, and collecting data relative to the updated reference model and relative to the orthodontic appliance.

Abstract: Disclosed herein are systems, methods, and devices for classifying ophthalmic images according to disease type, state, and stage. The disclosed invention details systems, methods, and devices to perform the aforementioned classification based on weighted-linkage of an ensemble of machine learning models. In some parts, each model is trained on a training data set and tested on a test dataset. In other parts, the models are ranked based on classification performance, and model weights are assigned based on model rank. To classify an ophthalmic image, that image is presented to each model of the ensemble for classification, yielding a probabilistic classification score—of each model. Using the model weights, a weighted-average of the individual model-generated probabilistic scores is computed and used for the classification.

Abstract: The invention relates to a device (1) and a method for determining the action of active ingredients on nematodes and other organisms in aqueous tests.

Abstract: The present invention relates to an apparatus for adaptive contouring of a body part. It is described to provide (210) at least one image; wherein, the at least one image comprises a first image comprising image data of a body part. An initial automatic model based segmentation of image data of the body part in the first image is determined (220). Final segmentation data of the body part is determined (230) in response to a modification of the initial automatic model based segmentation. An updated model based segmentation can be applied (240) on the basis of the initial automatic model based segmentation and the final segmentation data.

Abstract: A method for providing position information for controlling at least one function of an environment provided with an environment coordinate system comprises the steps of receiving image information associated with at least one image of at least one part of the environment captured by a camera of a mobile device, the mobile device being separate from the environment, determining a position associated with the mobile device in the environment coordinate system according to the image information associated with the at least one image, and providing position information indicative of the position associated with the mobile device in the environment coordinate system to control at least one function of the environment according to the position. There is also provided a method and system which provides movement information indicative of a movement associated with the mobile device relative to the environment coordinate system to control at least one function of the environment according to the movement.

Abstract: An image processing device is an image processing device configured to edit and output a video having been input. The image processing device includes a feature portion detecting unit configured to detect a feature portion from a frame incorporated in the video having been input, and a setting unit and an image generating unit configured to cause a playback speed of a frame incorporated in the video having been input and including the feature portion detected to be different from a playback speed of a frame including no feature portion detected, and output the video.

Abstract: A motion compensating prediction method includes: determining a location of an initial reference pixel of a current pixel in a reference image, where the current pixel is located in a first sub-image in a current image, when the initial reference pixel is located outside a second sub-image at a location corresponding to the first sub-image in the reference image, determining a location of a target reference pixel of the current pixel in the reference image based on the location of the initial reference pixel, where location precision of the target reference pixel is limited to being less than or equal to preset pixel location precision, and determining a predicted value of a pixel value of the current pixel based on a pixel value of the target reference pixel and/or a pixel value of a neighboring pixel of the target reference pixel.

Abstract: An example image processing device obtains first images capturing a street, on which a target vehicle is running, from a plurality of directions, and position information indicating positions at which the first images are taken, and associates a background object in a second image with a background object in a third image. The second image is an image which the target object is detected from and the third image is an image which is taken by an imaging device having a common imaging range with the second image among imaging devices mounted on the target vehicle. An output image is generated to include the position information indicating a position at which the second image is taken, the image of the detection rectangle which is clipped from the second image and includes the target object and the background object, and the first image of surroundings of the target vehicle.

Abstract: A transport service system automatically associates road sign information with road segments based on image data including a representation of the road sign. The transport service system receives image data from one or more camera systems. Based on the image data, the transport service system determines a distance score and an orientation score for one or more candidate road segments. The transport service system additionally determines whether the image data is anomalous. Based on the determined distance and orientation scores, the transport service system generates a composite score and a confidence score for each candidate road segment. The confidence score is adjusted based on whether the image data is anomalous. Based on the generated scores, the transport service system identifies a most likely candidate road segment for association with the road sign.

Abstract: A thumbnail-based image display method and terminal, where the method includes displaying a thumbnail on a screen, performing feature extraction on the thumbnail to obtain a feature factor of the thumbnail, displaying prompt information on the screen when the feature factor of the thumbnail is matched with a feature factor in a feature factor index library, where the feature factor index library indicates correspondences among feature factors and image storage addresses, obtaining an original image obtaining request, and displaying the original image on the screen based on the original image obtaining request. According to the method and terminal, feature indexes are created for all images in memories and buffers of all applications and in an image library, and an original high-resolution image is found using a feature factor of a thumbnail, thereby implementing high-definition image sharing between different applications.

Abstract: The present invention provides a machine-vision-based electronic automobile insurance fee meter. The machine-vision-based electronic automobile insurance fee meter can capture a lane departure behavior, and record a distance and duration of a lane departure in a memory module; also can capture a following-too-close behavior, and record a distance and duration of dangerous following in the memory module; and also can calculate a journey-based mileage insurance fee, a total mileage insurance fee P being a product of a mileage L of a journey, a unit mileage (per kilometer) insurance rate Rkm and a safety floating factor f. The present technical solution charges the insurance fee according to the actual driving mileage, makes the charging of the insurance fee fairer and more reasonable, and encourages safe driving and driving less; and also can identify the most important dangerous driving risks including following-too-close and lane departure, making the pricing of the insurance fee more accurate and reasonable.

Abstract: An imaged face image of a comparison target person is compared to a registered face image. When a comparison score indicating a result of the comparison is equal to or smaller than Th1 and it is determined that the comparison target person in the imaged face image is not a subject of the registered face image, it is determined whether or not the comparison score is equal to or greater than Th2 (Th2<Th1). When the comparison score is equal to or greater than Th2, a partial score indicating similarity between the imaged face image and a portion of the registered face image is calculated. When the calculated partial score is equal to or smaller than Th3, processing of emphasizing a portion corresponding to the partial score is performed on at least one of the imaged face image and the registered face image. The processed image is displayed.

Abstract: A method for generating a composite image includes receiving a first image that includes an eyeglass lens illuminated by a first illumination source radiating electromagnetic radiation in a first polarization state, and receiving a second image that includes the eyeglass lens illuminated by a second illumination source radiating electromagnetic radiation in a second polarization state. The second polarization state is different from the first polarization state, and the second illumination source is spatially separated from the first illumination source. The method also includes identifying, in the first image, a first portion that represents a reflection of the first illumination source on the eyeglass lens, and generating the composite image in which the first portion is replaced by a corresponding second portion from the second image. The first portion and the second portion represent substantially a same portion of the eyeglass lens.