282results about How to "Improve reconstruction effect" patented technology

The invention provides a video data storage method and system, which are applied to a distributed video monitoring system comprising a management node, storage nodes, a writing client and a reading client. The method comprises the steps that the management node receives a stripe resource request message from the writing client and allocates stripes to the writing client, wherein each stripe comprises at least one data block used for storing verifying blocks; the management node sends a data block resource request message to the storage nodes corresponding to the stripes and receives a data block resource response message carrying data block information and returned by each storage node; and the management node determines information of the stripes by utilizing the data block information and sends the information of the stripes to the writing client through a stripe resource response message, and the writing client writes video data into corresponding data blocks of the storage nodes corresponding to the stripes by utilizing the information of the stripes. Through the technical scheme provided by the invention, high writing performance, high space utilization rate and high reconstruction performance are supported.

The invention relates to an image encryption method and an image decryption method with visual security and data security based on compressed sensing. The image encryption method comprises the steps of: firstly, utilizing an SHA 256 hash function to obtain a 256-bit hash value of a plaintext image as an image secret key, and calculating initial numerical values of one-dimensional skew tent chaotic mapping and zigzag scrambling; carrying out sparse processing on the plaintext image, and carrying out zigzag scrambling on a coefficient matrix; and then utilizing the one-dimensional skew tent chaotic mapping to generate a measurement matrix, measuring and quantifying a scrambling matrix to obtain a compressed and encrypted image, and embedding the image into a carrier image with visual significance to obtain a final ciphertext image with visual significance. The image encryption method realizes the visual security and data security of the plaintext image, has large secret key space, is highly sensitive to plaintext, has higher capacity of resisting brute-force attack, chosen-plaintext attack and known-plaintext attack, does not need an additional storage space, and can transmit and store the ciphertext image quickly and effectively.

The invention discloses a 3D reconstruction method for weighing stereo matching and visual appearance, the specific steps of which are as follows: (1) making a sample database, including a stereo matching depth map, a visual appearance depth map, a multiview RGB chart and a real depth map; (2) constructing a depth convolution neural network; (3) using database to train neural network for the weight distribution value of the stereo matching and the visual appearance, the structure of the neural network is adjusted according to this value, until the neural network model is obtained with a better effect; (4) inputting the stereo matching depth map, the visual appearance depth map and the RGB chart, and a new depth map is acquired through the neural network model; (5) reconstructing a 3D model with the new depth map. The 3D reconstruction method combines the two practices of the stereo matching and the visual appearance, through the distribution weight value of the depth neural network, and can not only solve the problem that the stereo matching cannot rebuild a highlight and have no texture region, but also avoid the defect that the visual appearance cannot reconstruct a concave. The high quality reconstruction of the complex objects is realized.

The present invention relates to an image information encoding and decoding method and a device for same. One embodiment of an image information encoding method according to the present invention, as an image information encoding method according to another embodiment of the present invention, includes the steps of: generating a restore block; applying a deblocking filter on the restore block; applying a Sample Adaptive Offset (SAO) on the restore block having the deblocking filter applied thereon; and transmitting information on the SAO application. During the applying of the SAO, the SAO is applied to chroma pixels, and during the transmitting of the information, in addition to information on whether the SAO is applied on the chroma pixels, at least one of area information, division information on the SAO coverage area, SAO type information, and SAO offset information is transmitted.

The invention relates to a super-resolution image reconstruction method based on analysis sparse representation, belonging to the technical field of image processing. The method comprises the following steps of: performing dictionary training according to a training sample set; and training a high-resolution dictionary and a low-resolution dictionary for an extracted feature; converting an image to be input from an RGB (Red, Green and Blue) space into a 1 alpha beta space and dividing into blocks of a same size; performing two kinds of operation on the blocks, wherein one is that each block is amplified by using the conventional amplification method and the other one is that an residual image of each block is extracted, sparse representation of the residual image in the low-resolution dictionary is calculated, and then the residual image is reconstructed in the high-resolution dictionary to obtain a reconstructed residual image; summarizing results of the two steps, converting back into the RGB space and performing back projection to obtain the reconstructed super-resolution image. According to the method, the image reconstruction noise can be obviously reduced, and detail features are kept; and meanwhile, the method has the advantages of easiness in operation and wide application.

The invention discloses an image super-resolution method based on a generative adversarial network. The method comprises the following steps: obtaining a training data set and a verification data set;constructing an image super-resolution model, wherein the image super-resolution model comprises a generation network model and a discrimination network model; initializing weights of the establishedgenerative network model and the discriminant network model, initializing the network model, selecting an optimizer, and setting network training parameters; simultaneously training the generative network model and the discriminant network model by using a loss function until the generative network and the discriminant network reach Nash equilibrium; obtaining a test data set and inputting the test data set into the trained generative network model to generate a super-resolution image; and calculating a peak signal-to-noise ratio between the generated super-resolution image and a real high-resolution image, calculating an evaluation index of the image reconstruction quality of the generated image, and evaluating the reconstruction quality of the image. According to the method, the performance of reconstructing the super-resolution image by the network is improved by optimizing the network structure, and the problem of image super-resolution is solved.

The invention provides a method for three-dimensional reconstruction of a liver computed tomography (CT) image. The method comprises the steps of segmenting a liver two-dimensional CT image sequence, respectively extracting segmentation sequences corresponding to every tissue in a plurality of tissues of the liver; and performing three-dimensional image reconstruction on every tissue according to the segmentation sequences according to every tissue and three-dimensional reconstruction processes corresponding to every tissue so as to perform three-dimensional reconstruction on the liver CT image. According to the technical scheme, when the three-dimensional reconstruction is performed on the liver CT image, the corresponding reconstruction processes can be selected according to the characteristics of different tissues, and accordingly reconstruction efficiency of the liver CT image and quality of the liver CT three-dimensional reconstruction image are improved.

The invention relates to the technical field of image super-resolution reconstruction, in particular to an image reconstruction system and method based on a cross residual channel-spatial attention network. The system comprises a shallow feature extraction module, a depth feature extraction module, an up-sampling module and a reconstruction layer. The input of the shallow feature extraction moduleis a low-resolution image, and the shallow feature extraction module is used for extracting shallow features; the depth feature extraction module comprises a frequency division module and a cross residual group, the input of the depth feature extraction module is the output of the shallow feature module, and the depth feature extraction module is used for extracting deep features; the input of the up-sampling module is a deep feature and is used for up-sampling; and the reconstruction layer is used for reconstructing features to obtain a high-resolution image. The reconstruction network provided by the invention has stronger feature expression capability and differentiated learning capability, and can reconstruct a high-quality high-resolution image.

The invention discloses a compressive-sensing-based digital image watermark embedding and extraction method, and belongs to the technical fields of information hiding and image processing. The method comprises the following steps of performing sparsification processing on binary digital image watermark information to obtain sparse one-dimensional watermark information, and constructing a measurement matrix which is used as a key; measuring the one-dimensional watermark information to realize the compression and the encryption of the one-dimensional watermark information by using the key; performing discrete cosine transform on original carrier image information, blocking the transformed discrete cosine transform-domain information, and embedding the processed watermark information into a carrier image to obtain a watermark-containing image. According to the method, a compressive sensing method is used for processing watermarks, so that dual functions of watermark compression and encryption are realized, the number of the embedded watermarks is increased, the invisibility and the security are enhanced, and the overall performance of the watermarks is remarkably improved; an SBHE (scrambled block hadamard ensemble) matrix is applied, so that the method is small in memory space, high in calculation speed and simple in hardware; sparsity is introduced into a compressive sensing reconstruction process, so that the reconstruction accuracy and effects are greatly improved.

The invention relates to a method for reconstructing the super resolution of a sequence image POCS. The method mainly comprises the following steps of: establishing an image enhancement observation model ykz=Hkz+nk, wherein k is more than or equal to 1 and less than or equal to p, degenerating a high-resolution image by the established image enhancement observation model to obtain multiple low-resolution observation images, establishing an image POCS reconstructing target equation for each low-resolution observation image, and carrying out optimization solving on the image POCS reconstructing target equation by using an iterative algorithm to obtain the currently estimated simulation high-resolution reconstructing image. By adopting the method for reconstructing the super resolution of a sequence image POCS, the Gibbs effect of a reconstructing image can be effectively restricted, and the quality and visual effect of the reconstructing high-resolution image are improved.

The invention provides a dynamic magnetic resonance image reconstruction method and device of adaptive parameter learning, and the method comprises the steps of carrying out the reconstruction of theregularization terms in a CS-MRI model, and includes using DCT in the spatial domain and using TV in the temporal domain to remove redundancy from dynamic magnetic resonance images, and using a convolutional neural network to carry out the adaptive learning on a large number of parameters in CS-MRI, and establishing a magnetic resonance image reconstruction model; reconstructing the sample image through the established magnetic resonance reconstruction model to obtain a reconstructed image; calculating a difference value between the full-sampling image and the reconstructed image; and according to the difference value, updating parameters in the model, including DCT, a TV filtering operator, a regularization parameter and the like, by using a back propagation algorithm in the network. According to the magnetic resonance image reconstruction model established by the method, the height under-sampling image can be efficiently reconstructed to obtain the image with very high reconstructionprecision and reconstruction speed, so that the time of magnetic resonance scanning can be effectively shortened under the condition that the spatial resolution is not lost.

The invention discloses a three-dimensional microscopic imaging method and system based on a light field microscopic system. The method comprises the following steps: S1, in an optical system, obtaining a first point spread function of a three-dimensional sample from an object plane to a main camera sensor plane and a second point spread function from the object plane to a secondary camera sensorplane, and generating a first forward projection matrix and a second forward projection matrix according to the first point spread function and the second point spread function; s2, acquiring a lightfield intensity image shot by a primary camera and a high-resolution intensity image shot by a secondary camera of the three-dimensional sample in an optical system; and S3, performing three-dimensional reconstruction on the light field intensity image, the first forward projection matrix, the high-resolution intensity image and the second forward projection matrix through a preset algorithm to generate a three-dimensional reconstruction result of the three-dimensional sample. According to the method, the focal plane reconstruction signal-to-noise ratio is enhanced under the same iteration frequency by adding one path of acquisition light path, and the reconstruction effect of light field microscopic imaging is greatly improved.

The invention relates to an ultrasonic signal reestablishing method based on sparse data, and belongs to the technical field of compressive sensing. Firstly, a Gabor atom library serves as a sparse transformation domain of ultrasonic signals, nonlinear transformation is carried out on the ultrasonic signals by using a measurement matrix to obtain measurement values, and original signals are reestablished through an orthogonal matching pursuit algorithm according to the measurement values and a sensing matrix. According to the ultrasonic signal reestablishing method based on the sparse data, an artificial fish school algorithm is introduced to optimize the orthogonal matching pursuit algorithm, the times of inner product calculation can be reduced, and calculation speed can be improved. The ultrasonic signal reestablishing method based on the sparse data has the advantages of being high in reestablishing accuracy and computing speed, and can be applied to fields such as data compression, transmission and storage and high resolution imaging.

The invention discloses an image reconstruction method based on combination of motion estimation and super-resolution reconstruction, and the method comprises the following steps: (1) performing interpolation amplification on all low-resolution images; (2) performing motion estimation on the image to be reconstructed and the reference images respectively; (3) fusing the images, thus getting a fused image; and (4) performing denoising treatment on the fused image, thus getting the super-resolution reconstructed image. The motion estimation process and the super-resolution reconstruction process are combined together, and the relation between a plurality of the reference images with different low resolutions and a motion estimation parameter of the image to be reconstructed is utilized, so that the precision of the motion estimation parameter is simultaneously optimized during the super-resolution reconstruction process, and the super-resolution reconstruction effect is effectively enhanced. Therefore, the method can be widely applied in the field of satellite remote sensing of biomedical images, image repairing and the like.

The invention discloses a deep learning high under-sampling hyperpolarized gas lung MRI reconstruction method. The method comprises the following steps: constructing and constructing a hyperpolarizedgas lung MRI image training set; according to the method, the cascade CNN model is used, lung contour information is added into a loss function, an accurate reconstructed image can be obtained under the high under-sampling multiple, and the imaging speed is remarkably increased.

The invention discloses a variable-length input super-resolution video reconstruction method based on deep learning. The method comprises the following steps: constructing a training sample with a random length, and obtaining a training set; establishing a super-resolution video reconstruction network model, wherein the super-resolution video reconstruction network model comprises a feature extractor, a gradual alignment fusion module, a depth residual error module and a superposition module which are connected in sequence; training the super-resolution video reconstruction network model by adopting the training set to obtain a trained super-resolution video reconstruction network; and sequentially inputting to-be-processed videos into the trained super-resolution video reconstruction network for video reconstruction to obtain corresponding super-resolution reconstructed videos. According to the method, a gradual alignment fusion mechanism is adopted, alignment and fusion can be carried out frame by frame, and alignment operation only acts on two adjacent frames of images, so that the model can process a longer time sequence relationship, more adjacent video frames are used, that is to say, more scene information is contained in input, and the reconstruction effect can be effectively improved.

The invention discloses an image super-resolution reconstruction method of a generative adversarial network based on Gaussian coding feedback, and the method comprises the steps: carrying out the preprocessing of an ImageNet data set, and manufacturing a reconstruction data set in one-to-one correspondence with a low-resolution image and a high-resolution image; Constructing a generative adversarial network model for training, and introducing a Gaussian coding feedback network into the model; Sequentially inputting the data sets obtained in the step A into a generative adversarial network formodel training; And inputting the low-resolution image to be processed into a trained generative adversarial network to obtain a high-resolution image. A generative adversarial network is formed by constructing a generation network and a discrimination network, a Gaussian coding feedback loop is added between the generation network and the discrimination network, more information is added to the generation network to guide the generation network to carry out training, and important features are added by improving a sub-pixel convolutional layer structure, so that useless information is reduced, and the reconstruction effect is improved.

The invention discloses a hyper-spectral compression perception reconstruction method based on nonlocal total variation and low-rank sparsity, and mainly solves the problems in the prior art that reconstruction accuracy is low and the effect is poor after compressed sampling of hyper-spectral data. The hyper-spectral compression perception reconstruction method comprises the steps that 1. the hyper-spectral data are inputted and vectorized; 2. the vectorized hyper-spectral data are sampled so that sampling data are obtained; 3. initial reconstruction of the sampling data is performed; 4. the initially reconstructed data are clustered; 5. the sampling data are classified according to the type of image elements so that various types of sampling data are obtained; 6. a secondary reconstruction model is constructed; and 7. The secondary reconstruction model is solved according to various types of sampling data so that the optimal data of secondary reconstruction are obtained, and the data act as the final reconstruction data. The idea of nonlocal total variation and clustering is introduced on the basis of low-rank sparse reconstruction so that the hyper-spectral compression perception reconstruction method has advantages of high reconstruction accuracy and great effect and can be used for hyper-spectral data imaging.