Abstract: An information generation method and apparatus, an electronic device, and a computer readable medium. The method comprises: generating a target region set on the basis of an obtained first target picture (201); selecting a target region satisfying a preset condition from the target region set as a candidate region to obtain a candidate region set (202); generating a second target picture on the basis of the candidate region set (203); and generating picture related information on the basis of the second target picture (204). Thus, effective extraction for a target region and further classification for the target region are achieved, so that an identification result of an image region is finer, and then the subsequent image processing technology has a promotion space.

Abstract: A system for forecasting when solar flares will occur and identifying their possible locations creates a mask to analyze an image of a solar disk. The system also creates an intensity histogram using the solar disk. Additionally, the system stores a resulting full-disk image and a sub-image identifying possible flare signature sites. Further, the system applies a method to filter out false positive detections by using an equation that enhances true positive pre-flare signatures and minimizes false positive signatures.

Abstract: A method includes obtaining input image frames, including at least two captured using different capture conditions, separating color channels of each image frame, and generating at least one bad pixel map for each color channel of each image frame. Each bad pixel map is generated by identifying one or more outliers in pixel values in the color channel based on an intensity distribution of pixel values in an operation window within the color channel. The operation window has a window size based on the capture condition and/or local image content in the corresponding image frame. The method also includes combining the bad pixel maps and performing a morphological operation to refine the combined bad pixel map and using one or more coordinates of one or more bad pixels in the refined bad pixel map to update one or more pixel values of at least one image frame.

Abstract: A recurrent unit is proposed which, at each of a series of time steps receives a corresponding input vector and generates an output at the time step having at least one component for each of a two-dimensional array of pixels. The recurrent unit is configured, at each of the series of time steps except the first, to receive the output of the recurrent unit at the preceding time step, and to apply to the output of the recurrent unit at the preceding time step at least one convolution which depends on the input vector at the time step. The convolution further depends upon the output of the recurrent unit at the preceding time step. This convolution generates a warped dataset which has at least one component for each pixel of the array. The output of the recurrent unit at each time step is based on the warped dataset and the input vector.

Abstract: The present application relates to a cable detection method, a robot and a storage device. The method includes: acquiring an RGB image of a cable and a corresponding depth image; performing image segmentation on the RGB image, and merging pixel points in the RGB image, which are adjacent in position and have similar pixel features, into super-pixels; determining a target similarity between at least some of the super-pixels and the adjacent super-pixels, and detecting cables in the RGB image according to the target similarity between at least some of the super-pixels and the adjacent super-pixels; determining whether the cables are overlapped; and when it is detected that the cables are overlapped, determining an overlapping relationship between the detected cables according to the depth image.

Abstract: A training system and a training method for a human-presence detection model are provided. The training method includes: obtaining, by a distance sensor and a temperature sensor, a plurality of training samples; dividing the plurality of training samples into a first data cluster and a second data cluster according to whether or not an input interface of the host device is in a used state; marking the first data cluster as a human-presence cluster; selecting a sub-cluster from the second data cluster; marking the sub-cluster as a human-absence cluster; and performing machine learning for the human-presence cluster and the human-absence cluster to generate the human-presence detection model. The first data cluster corresponds to the used state. A center point of the sub-cluster corresponds to a maximum distance value and a minimum temperature value.

Abstract: A non-transitory computer-readable recording medium stores therein an estimation program that causes a computer to execute a process including, identifying a first person who uses a first cart from a first image acquired by capturing inside a store, generating skeleton information of the first person, acquiring, by using the skeleton information, a first space in which the first person grasps a grip part of the first cart, and estimating a first scale of the first person based on the first space and length information of the first cart.

Abstract: A data marking method includes obtaining a to-be-recognized image that includes to-be-recorded data, the to-be-recorded data being displayed by a display apparatus, recognizing the to-be-recorded data in the to-be-recognized image, detecting the data content displayed by the display apparatus according to the to-be-recorded data, and displaying a marking according to a detection result.

Abstract: A method and apparatus for processing an image. The method may include: acquiring a top view of a road; identifying a position of a lane line from the top view; cutting the top view into at least two areas, and determining, according to the position of the lane line in each area, a width of a lane in the each area and an average width of the lane in the top view; calculating a first perspective correction matrix by optimizing a first loss function, the first loss function being used to represent a difference between the width of the lane in the each area and the average width of the lane in the top view; and performing a lateral correction on the top view through the first perspective correction matrix to obtain a first corrected image.

Abstract: An image capturing device with a sensor, a memory, and a processor is provided. The sensor captures an image of at least one target. The memory stores a plurality of instructions. The processor obtains the plurality of instructions to perform the following steps: controlling the sensor to capture a reference image and a processed image; capturing a first bright region and a dark region from the reference image, and capturing a second bright region from the processed image; performing calculations on a first brightness value of the first bright region and a second brightness value of the second bright region respectively with at least two first brightness thresholds, to obtain a first low exposure compensation value and a second low exposure compensation value; and obtaining a high exposure compensation value according to comparisons between a third brightness value of the dark region and at least two second brightness thresholds.

Abstract: An image edge detection method for processing an image including multiple pixels includes performing a convolution operation on the image by a gradient operator in a first direction and a gradient operator in a second direction to obtain a first-direction gradient data and a second-direction gradient data, wherein the first direction is perpendicular to the second direction, performing a gradient statistical calculation within a neighboring area of a target pixel of the image according to the first-direction gradient data and the second-direction gradient data to obtain a gradient statistic, and determining an edge significance corresponding to the target pixel according to the gradient statistic.

Abstract: A document management system performs optical character recognition (OCR) on a high resolution version of an image of a document, obtaining a first set of text representative of the text of the document. The document management system applies a machine-learned model on a low-resolution version of the image of the document, producing a denoised image that is of a higher resolution than that input into the machine-learned model. The document management system performs OCR on the denoised image, obtaining a second set of text representative of the text of the document. The document management system compares the first and second sets of text in the form of an objective loss function. The document management system subsequently generates a feature vector from the comparison of the sets of text and retrains the machine-learned model with the generated feature vector.

Abstract: Systems and methods for evaluating a set of bounding boxes in a blended image are described. A system can include an integrated circuit configured to obtain expected bounding box data. The expected bounding box data can be based on coordinates data of an image. The integrated circuit can determine target coordinates based on the expected bounding box data. The integrated circuit can receive a blended image including a set of objects and a set of bounding boxes. The integrated circuit can extract pixel values located at the target coordinates in the blended image. The integrated circuit can identify an error relating to the set of bounding boxes based on the extracted pixel values and the expected bounding box data.

Abstract: A metal artifact reduction (MAR) method and system for reducing metal artifacts from X-ray images. The method is suitable for CT images generated with metallic elements exterior to a patient being scanned, for example, where external probes are applied to a patient, such as ultrasound probes. The method may be embodied in a computer algorithm for use, for example, in radiotherapy treatment planning and patient positioning procedures. In one application, the disclosed technique improves the dose delivery accuracy in ultrasound-guided cardiac radioablation, making this treatment modality a viable option for cardiac arrhythmias.