60 results about "Pixel value differencing" patented technology

Numerical image processing of two or more medical images to provide grayscale registration thereof is described, the numerical image processing algorithms being based at least in part on a model of medical image acquisition. The grayscale registered temporal images may then be displayed for visual comparison by a clinician and/or further processed by a computer-aided diagnosis (CAD) system for detection of medical abnormalities therein. A parametric method includes spatially registering two images and performing gray scale registration of the images. A parametric transform model, e.g., analog to analog, digital to digital, analog to digital, or digital to analog model, is selected based on the image acquisition method(s) of the images, i.e., digital or analog/film. Gray scale registration involves generating a joint pixel value histogram from the two images, statistically fitting parameters of the transform model to the joint histogram, generating a lookup table, and using the lookup table to transform and register pixel values of one image to the pixel values of the other image. The models take into account the most relevant image acquisition parameters that influence pixel value differences between images, e.g., tissue compression, incident radiation intensity, exposure time, film and digitizer characteristic curves for analog image, and digital detector response for digital image. The method facilitates temporal comparisons of medical images such as mammograms and/or comparisons of analog with digital images.

After a plurality of differently exposed images are obtained, and tone levels of the plurality of images are adjusted, image elements corresponding to changes of objects are detected based on pixel value differences between the plurality of images. Then, the numbers of blown-out highlight and shadow-detail loss image elements in the image elements corresponding to the changes of the objects are counted for the plurality of images. Then, an image in which the total of the numbers of image elements is smallest is selected. Furthermore, image elements corresponding to the changes of the objects in the selected image are used as those corresponding to image elements corresponding to the changes of the objects in an HDR image to be generated.

A method for reducing blocking artifacts in a video stream comprises receiving a picture of the video stream, wherein the picture includes a plurality of macroblocks and each of the plurality of macroblock includes four blocks, determining blocks with quantization parameters greater than a first threshold value in the picture, checking if block boundaries of the blocks are sharp and are real edges of objects in the picture according to pixel value differences between two adjacent pixels respectively located at both sides of the block boundaries, selecting filtering strengths of a de-blocking operation according to the pixel value differences when the block boundaries are sharp and are not real edges of the objects in the picture, and performing the de-blocking operation for two adjacent blocks at both sides of the block boundaries.

A fast loss less image compression system based on neighborhood comparisons compares pixel value differences with neighboring pixels and replaces such pixel values with the minimum of the differences. A marker is attached to a block of pixels, such that all the pixels in that block are compared with neighbors of one direction. The marker indicates how all of the pixels in that block are compared. Intermittent Huffman-tree construction is used such that one tree is used for several frames. Huffman coding is used to compress the resulting frame. A single Huffman-tree is constructed once every predetermined number of frames. The frequency of Huffman-tree construction can be performed according to the instantaneous availability of processor time to perform the construction. When more processing time is available, the Huffman-trees are computed more frequently. Such frequency variation can be implemented by using an input video frame buffer. If the buffer is a certain size, then processor time for Huffman-tree construction is available.

The present invention relates to the fields of motion detection, scene change detection and temporal domain noise reduction. Especially, the present invention relates to a device, a method and a software for detecting motion and/or scene change in a sequence of image frames. The device for detecting motion and/or scene change in a sequence of image frames comprises a difference generator adapted to calculate one or more differences between values of pixels of separate image frames, the pixel values comprising color information, and a discriminator for providing an indication whether the calculated one or more pixel value differences correspond to noise or correspond to image content based on the one or more pixel value differences.

The invention relates to the technical field of image processing, in particular to an edge detection method and system based on image multidimensional analysis. The method comprises the steps: detecting the edge of a target image through a Canny edge detection algorithm, obtaining a first edge image, carrying out the frequency domain analysis of the target image, obtaining an original spectrogram, processing the original spectrogram, and obtaining a corresponding pixel number difference index, a gray level histogram difference index, an adjacent pixel value difference index and a pixel value difference average value index; calculating according to the indexes to obtain an optimal high-frequency spectrogram selection index, and further determining an optimal high-frequency spectrogram according to the optimal high-frequency spectrogram selection index; and integrating the optimal high-frequency spectrogram and the first edge image to obtain a final edge image, so that the details of the image obtained by a Canny operator can be increased, the noise influence in high-frequency information can be reduced and eliminated, and the edge detection accuracy is greatly improved.

Systems and methods described herein enable a residual error image to be added and rendered in an existing HTML canvas using native primitives. A current image and a residual error image containing pixel value differences between the current image and an updated image are received. A positive residual error image and a negative error image are generated from the residual error image. The positive residual error image is added to the current image to generate a partially updated image, and the partially updated image is xored to generate an inverse image. The negative residual error image is added to the inverse image to generate an updated inverse image, and xoring the inverse image to revert the inverse image to generate the updated image.

The invention relates to the technical field of artificial intelligence, in particular to a structural part production process defect detection method based on image processing. According to the method, a ray image of a structural part is obtained through X rays, and a defect distribution diagram is constructed through defect areas on the ray image. A pixel point with the minimum pixel value in the defect connected domain on the defect distribution diagram is taken as a dark pixel point, and screening is carried out twice according to the dispersion degree of the dark pixel point and the distance relationship between the dark pixel point and the center of the connected domain to obtain the defect connected domain with the defect type of slag inclusion and a second connected domain to be detected needing to be continuously detected. Twice screening is carried out through the first pixel distribution and the second pixel distribution of the pixel points in the second to-be-detected connected domain, and defect type detection of all defect connected domains is completed. According to the method, the pixel value difference characteristics of air holes and slag inclusions are fully considered, so that defect detection is accurately and efficiently completed.

The invention discloses a target detection method and device, electronic equipment and a storage medium. The method comprises the steps of determining at least one to-be-detected target object in a to-be-detected image, and obtaining a target template image of each target object; determining the image complexity of each target template image based on the pixel value difference of each pixel point in each target template image; determining a detection threshold corresponding to each target object according to the image complexity; recognizing the to-be-detected image according to the target template image to obtain a plurality of candidate object detection areas of each target object; for each target object, determining a primary object detection area of the target object from the candidate object detection areas according to the similarity between the candidate object detection areas and the target template image and a detection threshold; and determining a target object detection area of the target object from the primarily selected object detection areas. The target detection method and device can adapt to detection scenes with various types of targets, and the target detection accuracy is improved.

Embodiments relate to biasing an image noise filter to reduce edge and texture blurring of image data. Pixel values used to determine photometric coefficients for a bilateral filter are modified by offset values. The offset value for a pixel value is determined by applying a high pass filter to the pixel (referred to as the center pixel) and neighboring pixels of the center pixel. By adding the offset value to the center pixel value, the pixel value difference between the neighboring pixels and the center pixel becomes smaller for pixels on the same side of an edge as the center pixel. Thus, pixels on the same side of the edge get more weight in the bilateral noise filter. Conversely, pixels on the opposite side of the edge as the center pixel get less weight in the bilateral filter. As a result, the biased bilateral filter reduces blurring of edges and increases preservation of texture in the image data.

The invention discloses a display screen uniformity test method, which comprises the following steps of obtaining a display state image of a display screen to be tested; determining the display state image into an initial image; performing filtering processing on the initial image so as to generate a filtering image; generating a comparison image used for showing the interference pixel value of the initial image according to the pixel value difference of each pixel point on the initial image and the filtering image; generating a uniformity distribution image of the initial image according to the pixel value differences of each pixel point on the initial image and the comparison image. The invention also discloses a display screen uniformity test system. The method and the device provided by the invention aim at reflecting the uniformity distribution condition of the display screen through the test results.

The invention relates to an interface testing method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a screenshot of a to-be-tested interface, obtaining a pixel value difference between the screenshot and an expected image of the to-be-tested interface, and determining an interface testing result of the to-be-tested interface according to the pixel value difference between the screenshot and the expected image. Whether the interface is normally displayed or not can be determined through the pixel value difference between the screenshot and the expected image, and codes do not need to be written. Meanwhile, the interface testing method provided by the embodiment of the invention can be suitable for all page structures, so that codes do not need to be rewritten for different page structures, and the UI testing efficiency can be relatively high.

The invention discloses a method and device for calculating auto-focus statistical information. The method comprises the following steps: obtaining an image block that takes a pixel to be processed as a center pixel as a first image block, wherein the first image block is selected from photographed images and has the auto-focus statistical information to be extracted; according to the color of the center pixel of the first image block, selecting the preset number of sub-image blocks to form a second image block by taking the pixel which has the same channel as the center pixel and can form the sub-image blocks with preset sizes in the range of the first image block as the center; determining the strong edge contained in the first image block according to the pixel value difference information between the sub-image blocks in the second image block; and according to the strong edge contained in the first image block, calculating to obtain the auto-focus statistical information of the first image block by adopting a corresponding operator matrix. By adopting the above scheme, the accuracy of the auto-focus statistical information of the photographed objects can be increased.

The invention discloses a compression method of a 3D printing file. The method comprises the following steps of (1) traversing the image data of all layers, setting the data of which the pixel value is smaller than a set value as 0, and storing the new image data; (2) traversing the image data of all layers, obtaining a pixel value difference chart of the current layer and the previous layer, and judging and marking each layer as a difference layer, a non-difference layer or a reference layer; (3) classifying all data into valid data, gray data and invalid data according to the pixel values; (4) if the previous layer is the reference layer, classifying and coding all data of the layer; if the current layer is the difference layer, classifying and coding a difference graph of the layer; if the current layer is the non-difference layer, recording the number of layers; and (5) repeating the step (4) until all the layers are coded or the number of the layers is recorded. According to the method, the storage size of the 3D printing file is effectively compressed in a mode of invalid data filtering, subtraction of the adjacent layer image data and classification and coding.