Abstract: A method and a device for encoding/decoding images are disclosed. The method for encoding images comprises the steps of: deriving a scan type of a residual signal for a current block according to whether or not the current block is a transform skip block; and applying the scan type to the residual signal for the current block, wherein the transform skip block is a block to which transform for the current block is not applied and is specified on the basis of information indicating whether or not transform for the current block is to be applied.

Abstract: There is provided an image processing apparatus that performs noise reduction processing by using a patch obtained by dividing an image into small areas. The image processing apparatus includes a setting unit configured to set, for the image, a patch of interest, a search unit configured to search the image for a similar patch based on the patch of interest, a generation unit configured to generate one or more synthesis patches different from a patch included in the image, based on at least one of the patch of interest and the similar patch, and a noise reduction unit configured to reduce noise in the patch of interest by using a patch group including the patch of interest, the similar patch, and the one or more synthesis patches.

Abstract: There are provided an image processor, an image processing method, and a program that acquire visible pixel information regarding a pixel viewed by a user in an image and perform blur processing on a basis of the visible pixel information and depth information indicating a depth value corresponding to each of pixels of the image during a predetermined period after the visible pixel information is acquired, thereby further enhancing a sense of immersion as if in a different space.

Abstract: An image processing method includes: grouping input image data of a dilation convolution to obtain Dh×Dw grouped image data; wherein Dh is a dilation rate of a convolution kernel corresponding to the dilation convolution on a height dimension thereof, Dw is a dilation rate of a convolution kernel corresponding to the dilation convolution on a width dimension thereof, and both Dh and Dw are positive integers; performing a convolution calculation on the Dh×Dw grouped image data respectively by a first convolution kernel, to obtain Dh×Dw grouped convolution calculation results; wherein the first convolution kernel is a convolution kernel before the dilation convolution is dilated; and obtaining a dilation convolution calculation result of the input image data according to the Dh×Dw grouped convolution calculation results. The present disclosure can reduce power consumption, improve efficiency of the dilation convolution calculation.

Abstract: An operation method of an image sensor includes detecting a motion region based on a first image and a second image to obtain a detected motion region, the first image corresponding to a first exposure time and the second image corresponding to a second exposure time, the second exposure time being shorter than the first exposure time, determining a weight for the detected motion region, performing signal pre-processing on the first image to generate a pre-processed image, and outputting a third image based on the pre-processed image and the weight.

Abstract: A signal to noise ratio adjustment circuit is configured to determine, whether a signal to noise ratio of a first image is below a first threshold and to determine, whether a variation of imaged content between the first image and a preceding second image of a video sequence is below a second threshold. The signal to noise ratio adjustment circuit is further configured to generate a third image having an increased signal to noise ratio as compared to the first image or the second image if the signal to noise ratio is below the first threshold and if the variation is below the second threshold.

Abstract: The present disclosure relates to an artificial intelligence (AI) system utilizing a machine learning algorithm, including deep learning and the like, and application thereof. In particular, an electronic device of the present disclosure comprises: a memory including at least one command; and a processor connected to the memory so as to control the electronic device, wherein, by executing the at least one command, the processor acquires an image, acquires a noise correction map for correction of noise of the image on the basis of configuration information of a camera having captured the image or brightness information of the image, and eliminates the noise of the image through the noise correction map. In particular, at least a part of an image processing method may use an artificial intelligence model having been acquired through learning according to at least one of a machine learning algorithm, a neural network algorithm, and a deep learning algorithm.

Abstract: An ultrasonic diagnostic imaging system acquires received beams of echo signals produced in response to a plurality of transmit events. The received beams are combined with refocusing to account for differences in receive beam to transmit event locations. The delays and weights used in the refocusing are supplemented with delays and weights which correct for reverberation artifacts. The received echo signals are processed to detect the presence of reverberation artifacts and a simulated transmission of reverberation signal components to virtual point sources in the image field is calculated. This simulation produces the delays and weights used for reverberation signal compensation, or estimated reverberation signals which can be subtracted from received echo signals to reduce reverberation artifacts.

Abstract: An imaging system (500) includes a data acquisition system (515) configured to produce projection data and at least one memory device with reconstruction algorithms (518) and at least one blending algorithm (524). The imaging system further includes a reconstructor (516) configured to reconstruct the projection data with the reconstruction algorithms and generate at least first spectral volumetric image data corresponding to a first basis material content and second spectral volumetric image data corresponding to a second basis material content, and blend the first spectral volumetric image data and the second spectral volumetric image data with the at least one blending algorithm to produce blended volumetric image data.

Abstract: A method for image selection applied to a mobile terminal including a camera may include: acquiring image information of a first image, the image information of the first image comprising a first acceleration value of the mobile terminal in response to the camera capturing the first image, the first image being any one of at least two images successively captured by the camera; and determining that the first image satisfies a multi-frame denoising processing condition in response to the first acceleration value being smaller than a preset acceleration threshold.

Abstract: In order to improve the noise suppression in a video image stream 3 of a medical image recording system, the video image stream including a sequence of frames, it is provided that an image processing unit 5 of the image recording system analyses the video image stream 3 continuously in real time and determines at least one variability between successive image pixels of the frames, for example of spatially adjacent image pixels of frames and/or of image pixels of a plurality of the frames corresponding to one another spatially and temporally, in order, on the basis of the variability determined, to set at least one parameter of a noise suppression subsequently applied to the video image stream 3. As a result, the noise suppression can be adapted continuously to a current recording situation.

Abstract: The present invention discloses an image adaptive noise reduction method and a device thereof. The method includes: dividing an original image into a plurality of sub-blocks; performing a space conversion for all the sub-blocks; performing a significance analysis to obtain a significant characteristic map; performing a threshold segmentation on all the significant characteristic maps by a significant standard value to obtain a significant characteristic region and a non-significant characteristic region; performing adaptive filtering on the significant characteristic region and maxing an original image in the non-significant characteristic region to obtain a mixed image; and performing an image space inverse conversion for the mixed image, and outputting a final image. The present invention uses the method of dividing the image, based on the significant characteristics of the image, reducing noise reduction in the non-significant characteristic region, saving algorithm running time and hardware resources.

Abstract: An image processing apparatus includes an acquisition unit configured to acquire a parameter about atmospheric fluctuation, a determination unit configured to determine a method and condition for noise reduction processing, a processing unit configured to perform, on input image data, noise reduction processing corresponding to the determined method and condition for noise reduction processing, and a generation unit configured to generate image data corrected by a speckle imaging method, based on a plurality of pieces of image data on which the noise reduction processing has been performed. The determination unit determines the method and condition for noise reduction processing based on the parameter.

Abstract: Embodiments relate to lateral chromatic aberration (LCA) recovery of raw image data generated by image sensors. A chromatic aberration recovery circuit performs chromatic aberration recovery on the raw image data to correct the resulting LCA in the full color images using pre-calculated offset values of a subset of colors of pixels.

Abstract: The present disclosure is related to systems and methods for noise reduction. The method includes determining a first image block in each of a group of continuous frames including a current frame and a plurality of reference frames based on coordinates of a target pixel in the current frame. The method includes determining at least one second image block corresponding to the first image block in each of the group of continuous frames. The method includes determining an average offset between the at least one second image block in the current frame and the second image blocks in the plurality of reference frames. The method includes determining a spatial domain filtering intensity and a temporal domain filtering intensity for the target pixel in the current frame.

Abstract: Techniques for lossless compression of a digital image using prior image context.

Abstract: In some aspects, one or more processors may perform an analysis of a processed image using an artificial intelligence module. Based on the analysis, the one or more processors may determine that the processed image includes one or more threats and determine details associated with individual threats of the one or more threats. The one or more processors may determine, based on the details associated with the one or more threats, that a particular threat threshold of a plurality of threat thresholds has been satisfied. The one or more processors may add one or more annotations to the image to create an annotated image that includes the one or more threats and at least a portion of the details associated with individual threats of the one or more threats. The one or more processors may send a notification to one or more designated recipients.

Abstract: An image processing device includes: a memory; and a processor comprising hardware, wherein the processor is configured to: receive agent observation image information including information of a plurality of pixels obtained by capturing an image based on fluorescence from a subject administered with an agent that emits fluorescence upon being irradiated with excitation light in a predetermined wavelength band; amplify pixel values of the plurality of pixels by executing first gain processing on the agent observation image information; and reduce pixel values of pixels lower than a predetermined threshold by executing reduction processing on the agent observation image information after the first gain processing.

Abstract: A method may include acquiring MR signals by an MR scanner and generating image data in a k-space according to the MR signals. The method may also include classifying the image data into a plurality of phases. Each of the plurality of phases may have a first count of spokes. A spoke may be defined by a trajectory for filling the k-space. The method may also include classifying the plurality of phases of the image data into a plurality of groups and determining reference images based on the plurality of groups. Each of the reference images may correspond to the at least one of the phases of the image data. The method may further include reconstructing an image sequence based on the reference images and the plurality of phases of the image data.

Abstract: An image processing apparatus obtains a first output image by applying an image to a first training network model, obtains a second output image by applying the image to a second training network model, and obtains a reconstructed image based on the first output image and the second output image. The first training network model is a model that uses a fixed parameter obtained through training of a plurality of sample images, the second training network model is trained to minimize a difference between a target image corresponding to the image and the reconstructed image.