45 results about "Texture coding" patented technology

An apparatus and method is provided for highly scalable intraframe video coding. The conventional macroblock DCT tools are integrated with the subband filter banks for the improved efficiency of scalable compression. The enhancement layers are represented in a subband domain and coded by an inter-layer frame texture coder utilizing inter-layer prediction signal formed by the decoded previous layer. Each quality enhancement layer is additionally scalable in resolution.

A texture encoding apparatus includes a texture data acquisition unit configured to acquire texture data of a texture set provided under a plurality of different conditions, a block segmentation unit configured to segment the texture data into a plurality of block data items each of which contains a plurality of pixel data items whose values corresponding to the conditions fall within a first range and whose pixel positions fall within a second range in the texture set, a block data encoding unit configured to encode each of the block data items to produce a plurality of encoded block data items, and a block data concatenation unit configured to concatenate the encoded block data items to generate an encoded data item of the texture set.

The present invention discloses a texture coding label for certificating authentification of a genuine commodity, wherein the texture coding label comprises a fabric texture portion having a plurality of fiber threads randomly distributed therein to form a fiber image, and a two-dimensional bar code portion for recording a two-dimensional bar code generated by calculating the fiber image and a serial number of the commodity through a first predetermined algorithm. Since the texture coding label combines the texture coding technology, two-dimensional bar code technology and digital encryption technology, a uniqueness could be ensure for the commodity so as to enhance the anti-counterfeiting function.

Methods of adaptive transform type based on transform unit (TU) size for enhancement layer (EL) coding and multiple motion candidates for EL coding based on corresponding base layer (BL) video data are provided. One method selects a transform type from multiple allowable transform types based on the TU size and applies the selected transform type to the transform units of the inter-layer prediction processed data. Another method derives multiple motion candidates for the EL video data coded in Merge mode or Inter mode based on motion information associated with the corresponding BL video data.

The invention provides a hand vein recognition method based on the fusion of structure coding characteristics and texture coding characteristics, belonging to the technical field of intelligent monitoring in computer vision. The method comprises the following steps of: step 1, acquiring and preprocessing an image; step 2, extracting the structure coding characteristics; step 3, extracting the hand vein texture coding characteristics; step 4, fusing the structure coding characteristics with the hand vein texture coding characteristics; and step 5, performing recognition through a sorter, thereby obtaining a result. The invention provides the hand vein recognition method based on the fusion of the structure coding characteristics and the texture coding characteristics, wherein binary coding is performed on the extracted structure characteristics and texture characteristics, which is advantageous for the retention of information in the characteristic fusion; the result obtained by fusing the characteristics is far better than the result of recognition obtained by only employing the structure coding characteristics and the result of recognition obtained by only employing the texture coding characteristics; and the robustness for image distortion and wrong segmentation is high, therefore, the hand vein can be correctly recognized in the presence of certain image distortion and wrong segmentation.

The invention discloses a method for detecting image characters in natural scene and relates to a method for adopting a texture descriptor LHBP to describe the texture character of an image and adopting a multi-dimension tropistic wave filtering method to detect characters in the image, thereby solving the problem that the character detection method based on texture has complex requirement on illumination; and the change of contrast gradient between the character and the background has great influence on detection. The obtained LHBP tropistic texture code and the corresponding code produced according to the change of position weight are obtained; a character region is determined through a multi-dimension tropistic analytic method. The method adopts a mode of extracting local texture on multi-dimension wavelet character by the LHBP texture descriptor, can filter out the influence of complexity and the change situation of the contrast ratio between the character and the background, effectively extracts the texture character of the character region, utilizes the texture direction property of the character region to determine the final character region and has good robustness in complex illumination, the change of the contrast gradient between the character and the background, the change of the size and the stroke thickness of the character and the like.

The invention discloses an image processing method of a coded pulse sequence. The image processing method comprises the following steps: performing multi-layer information coding on an image, and coding information of different layers of the image into a pulse sequence according to a time sequence relationship; converting an image into a grayscale image and standardizing the grayscale value of theimage; encoding the shape, the feature points, the color and the texture of the grayscale image into pulses; and arranging the coded pulses according to a set sequence to form a string of pulse sequence. According to the invention, the information of each layer carried by the image is effectively utilized, and meanwhile, the non-key information is reduced. For the spiking neural network, the information completeness of the pulse sequence input by the neural network is improved, and the redundancy of information is reduced, and the information processing efficiency of the spiking neural network can be improved. For signal processing, the information capacity of an information source and the coding efficiency of information are improved.

The invention discloses a texture processing method applied to a piece of electronic equipment including a first processing unit and a second processing unit. The method comprises: the first processing unit transmits a first texture having a first coding format to the second processing unit, wherein the first coding format does not belong to a texture coding format supported by the electronic equipment; the second processing unit decodes the first texture to obtain corresponding first texture color data; and the second processing unit codes the first texture color data to obtain a second texture having a second coding format that is a texture coding format supported by the electronic equipment. In addition, the invention also provides electronic equipment for texture processing. With the method and electronic equipment, the texture for cross-platform code usage is realized under the circumstances that the CPU resource consumption is reduced and the operating speed is increased.

The present invention relates to encoding a decoding video employing texture coding. In particular, a texture region is identified within a video picture and a texture patch is determined for said region. Clustering is performed to identify a texture region within the video image. The clustering is further refined. In particular, one or more brightness parameters of a polynomial is determined by fitting the polynomial to the identified texture region. In the identified texture region, samples are detected with a distance to the fitted polynomial exceeding a first threshold and identify a refined texture region as the texture region excluding one or more of the detected samples. Finally, the refined texture region is encoded separately from portions of the video image not belonging to the refined texture region.

The embodiment of the invention discloses a face image processing method and device, image equipment and a storage medium. The face image processing method comprises the steps of obtaining first key point information of a first face image; performing position transformation on the first key point information to obtain second key point information conforming to a second human face geometric attribute; wherein the second face geometric attribute is different from a first face geometric attribute corresponding to the first key point information; and performing face texture coding processing by using a neural network and the second key point information to obtain a second face image.

More efficient global motion compensation is provided by merging the warping processes performed in various global motion compensation (GMC) functions into a single warping process for use in GMC encoding and decoding operations. In an encoder in accordance with the invention, a global motion estimation processor is provided for performing global motion estimation (GME) on a picture. A mode decision processor is provided for selecting macroblocks (MBs) of the picture for GMC coding. A warping processor is provided for performing a single warping process for each pixel in the selected MBs for use in a plurality of GMC operations other than GME. A global motion compensation processor is provided for performing the GMC operations. Variable length encoding and texture coding is provided for encoding the picture to produce a GMC encoded bit stream.

The invention discloses an infrared image volume cloud detection method based on neighborhood intensity texture coding. The invention relates to the technical field of infrared image processing. The method comprises the following steps: firstly, creating a sliding window to traverse infrared image pixel points; performing texture coding on the infrared image by using a neighborhood intensity coding method to generate a texture image, roughly determining coordinates of a suspected point region based on local probability distribution, then using region growth to aggregate false alarm source pixels, and finally performing denoising by using a neighborhood statistical method to finally obtain a detection result. According to the method, the problems of difficulty in accurately detecting the volume cloud, low detection precision and fuzzy detection result in the existing infrared imaging technology are solved, and the method has the advantage of accurately aggregating the false alarm sourcepixel region.

The invention discloses an infrared imaging river channel detection method based on neighborhood intensity texture coding. The invention relates to the technical field of infrared image processing. The method comprises the following steps of: performing texture coding on an input image by using a neighborhood intensity coding method; generating a texture image, roughly determining coordinates of asuspected point region based on local probability distribution, generating a preliminary detection result by using region growth, neighborhood statistical denoising and other methods, detecting a connected region of the preliminary detection result, taking gradient features of the input image, and performing region growth to finally obtain a detection result. According to the method, the problemsthat the riverway is difficult to accurately detect, the detection precision is low and the detection result is fuzzy in the existing infrared imaging technology are solved, and the method has the advantage of accurately aggregating the false alarm source pixel region.

The invention discloses a three-dimensional head model reconstruction method and device based on a deformable neural radiation field, and the method comprises the steps: carrying out the segmentation and face key point detection of an input video frame by frame, and fitting a parameterized model frame by frame; semantic information is extracted from an input video frame by frame, wherein the semantic information mainly comprises a hair label and a face label; under the guidance of a parameterization model and semantic information, rigid registration coding, non-rigid deformation coding, texture coding, a deformation quantity estimation model, a topology estimation model, a signed distance field estimation model, a color estimation model and a semantic information estimation model are optimized on an input video in a deformable neural radiation field rendering mode; therefore, a frame-by-frame high-quality three-dimensional head model is obtained.

A signature fingerprint identification method comprises the following steps: 1) registration of a signature fingerprint and a rolling fingerprint and manufacturing of a weak label: registering the rolling fingerprint with the signature fingerprint, and extracting a direction field and minutia points of the rolling fingerprint by using a traditional algorithm to serve as the weak label of the signature fingerprint; 2) training a multi-task full convolutional neural network based on signature fingerprint image enhancement of a weak label and minutiae extraction, taking the weak label of the signature fingerprint as a training label, and obtaining a full convolutional neural network model which simultaneously generates a signature fingerprint enhancement image and minutiae; and 3) performing multi-score strategy fusion based on the minutiae template and the texture template, performing minutiae coding and texture coding on the generated signature fingerprint enhanced graph and the minutiae, performing comparison, and performing strategy fusion on the compared scores to obtain the final comparison score of the fingerprint. According to the invention, specialists do not need to manually mark fingerprint minutia points, a large amount of time and labor cost are saved, and the comparison accuracy is high.

A motion evaluating device and method with optimized complication degree. Disclosed is a motion evaluating (ME) device and method using a motion vector (MV) fast searching algorithm, which can significantly reduce calculation amount by using the fast searching algorithm rather than a traditional global search algorithm, and which will allows the user to define the complexity level of the calculation. The MV fast search algorithm induces about 0.5 dB of image degradation at most, but can significantly reduce the calculation amount. As the calculation amount of the ME circuit can be adjusted by inputing the calculation complexity parameters, it make a pipeline arrangement taking a parallel approach to implement the texture coding in it, have a highest efficiency.

The invention discloses a texture coding image-based three-dimensional reconstruction method, and the method comprises the following steps: 1, projecting a color coding image generated through the coding of an M array to a to-be-measured object, and shooting a to-be-measured region through a camera, and obtaining a left image and a right image; and 2, extracting a target area based on a Hough circle detection method and a perspective transformation principle so as to avoid the influence of environmental sundries on a projection area. Step 3, performing image enhancement and color identification preprocessing based on a color migration technology, and converting color information of the image into a code value; and step 4, directly decoding for an M array coding method to obtain matching point pairs. And step 5, performing three-dimensional reconstruction based on a stereoscopic vision principle by using the obtained homonymy point pairs to obtain three-dimensional coordinates of the spatial points corresponding to the two-dimensional points. According to the invention, space coding and decoding and binocular vision structured light three-dimensional reconstruction technologies are combined, the three-dimensional reconstruction speed and precision are improved, and a new thought is provided for high-efficiency three-dimensional reconstruction on the premise that the high simulation degree is kept.

The invention provides a CU segmentation prediction and mode decision texture coding method based on a JND model. The method comprises the following steps: firstly starting a mode decision, and constructing a multi-view texture JND model of a texture video through a brightness JND model, a space JND model and a time JND model; secondly, setting a threshold value of a multi-view texture JND model according to the content of the texture video, and dividing tree blocks of the texture video into three types according to the threshold value; performing adaptive CU segmentation on the current tree block according to the type of the current tree block; and finally, according to the type to which the current tree block belongs, performing fast mode decision prediction on the tree block segmented by the adaptive CU, and determining the optimal coding mode of the tree block. The JND model is used to analyze the tree block features of the texture image, some tree blocks of the texture video are skipped in the early stage, the coding complexity of 3D-HEVC can be significantly reduced, and the loss of the RD performance of the video can be ignored.

The present invention discloses a multiview video coding and decoding method and device, and a related device. The multiview video coding and decoding method comprises the steps of: determining abandoned depth map frames for a depth map sequence in an original multiview video; performing compression coding of texture video frames and non-abandoned depth map frames to obtain texture coding data andnon-abandoned depth map frame coding data; employing the texture coding data and the non-abandoned depth map frame coding data to perform decoding reconstruction to obtain a reconstructed depth map sequence; determining a filtering coefficient according to an original depth map sequence and the reconstructed depth map sequence; and sending the texture coding data, the non-abandoned depth map frame coding data and the filtering coefficient to a decoding terminal. Part of the depth map frames is abandoned, so data requiring coding is reduced, and therefore the coding efficiency of multiview video is improved, and virtual viewpoints and views with high quality are obtained.

At least one of the embodiment provides a method and device for generating and encoding in a bitstream an interpolation texture coding mode indicating that the bitstream contains color information data representative of a texture image and that texture interpolation has to be done on point of a reconstructed point cloud that are not colorized from said color information data.

Regions for texture-based coding are identified using a spatial segmentation and a motion flow segmentation. For frames of a group of frames in a video sequence, a frame is segmented using a first classifier into a texture region, a non-texture region or both of an image in the frame. Then, the texture regions of the group of frames are segmented using a second classifier into a texture coding region or a non-texture coding region. The second classifier uses motion across the group of frames as input. Each of the classifiers is generated using a machine-learning process. Blocks of the non-texture region and the non-texture coding region of the current frame are coded using a block-based coding technique, while blocks of the texture coding region are coded using a coding technique that is other than the block-based coding technique.