31results about How to "Guaranteed reversibility" patented technology

The invention discloses a digital audio reversible watermarking algorithm based on side information prediction and histogram shifting. The algorithm includes the following steps that the prediction coefficient of audio data and the prediction value of a sample are calculated; the prediction value is subtracted from the sample value to obtain a prediction error, two different histograms are constructed according to two modes of the side information prediction algorithm, and then watermarking information is embedded in the prediction error to obtain the prediction error containing a watermark; the prediction error containing the watermark is embedded in a next sample value of the current sample to obtain an audio value containing the watermark; an audio signal containing the watermark is received at the receiving end, the prediction value is subtracted from the sample value, an embedded signal is extracted by means of prediction error histogram shifting characteristics, and an original audio signal is restored in a lossless mode. Compared with an original prediction algorithm, the algorithm improves prediction performance; non-causality and histogram shifting are combined, so the algorithm has good embedability and low distortion rate when the embedding rate is low.

The invention relates to a reversible video frequency watermark method based on interframe forecast error histogram modification, comprising watermark embedding and watermark detection. The watermark embedding process is carried out according to a video frequency negative sequence: firstly, utilizing the previous frame to carry out motion estimation to obtain an interframe forecast error, and then carrying out the histogram modification on the interframe forecast error so as to embed a watermark. An extracting process is carried out according to a video frequency sequence mode: similarly, utilizing the previous frame to carry out the motion estimation to obtain a forecast error, and then extracting a watermark from the forecast error, and modifying a histogram to carry out the frame imagerecovery. The invention uses the mode of negative sequence embedding and sequence extracting, thereby an embedded video frequency frame can be recovered after the watermark is extracted, and the reversibility of the video frequency watermark is realized; the motion estimation is used according to the content relation of adjacent interframes to carry out prediction so as to enable the forecast error histogram to have the height centrality, and realize the bigger embedding content; in the embedding process, using the histogram modification enables that the influence on the frame image is small, and a watermark-embedded video frequency has high visual quality.

The present invention relates to the fields of digital forensics and information security, and more specifically, relates to a binary image reversible information hiding method based on an image enlargement strategy, which can restore the original image losslessly after information extraction. Use the cross-shaped pattern block to enlarge the original image to achieve the purpose of reversible information hiding. By analyzing and classifying 32 kinds of cross-shaped pattern blocks, and selecting the appropriate pattern block as the reference pattern block, a reference image with better visual quality can be obtained. ; After determining the reference pattern block, analyze the type of cross-shaped pattern block centered on each pixel in the original image, and use the designed enlargement strategy to enlarge each pixel into a 2×2 block to obtain the reference image; Finally, a selection strategy of flippable pixels in a block is adopted to flip the pixels to embed information. The invention can be used in information security, especially in the transmission of binary image secret information.

The present application relates to a sodium metal battery and an electrochemical device. The battery includes a positive pole piece and a negative pole piece. The negative pole piece is a negative current collector. The thickness of the deposited sodium layer is ≥30nm. The present invention utilizes the first irreversible capacity of the positive electrode material and the design optimization of the battery cell. After the battery cell is charged and discharged for the first time, the amount of residual sodium metal is large enough to uniformly form a layer of sodium deposition layer with a certain thickness on the surface of the negative electrode current collector. Avoid the higher nucleation energy required for sodium to deposit on the surface of the current collector during the subsequent charge-discharge cycle, reduce the overall deposition overpotential, and ensure the uniformity of sodium metal deposition and the reversibility of the charge-discharge process.

An electric actuator, in particular for driving a gearbox selector shaft, the actuator comprising a coil guided in rotation between two flat permanent magnets carried by yoke plates, the coil being centered on the rotary spindle and having an angular stroke lying in the range 90° to 180°, its rotary spindle being connected by stepdown gearing to the gearbox selector shaft.

The invention discloses a high-fidelity H.264 / AVC video three-coefficient reversible steganography method which comprises two parts of an embedding method and an extraction method. The embedding method comprises the steps of: according to an H.264 / AVC video standard, carrying out coding operation on a YUV video sequence frame by frame to obtain quantized residual DCT coefficients of all macro blocks in a current coded frame; sequentially traversing a plurality of macro blocks in the current coded frame according to a sequence from top to bottom and from left to right, and selecting 4*4TU sub-blocks of which texture-complexity is greater than or equal to a set threshold; and according to a three-coefficient embedding strategy, completing embedding, and carrying out entropy coding to obtaina compressed video after embedding. According to the invention, a three-coefficient modulation strategy combines information of a vacuum value, a coefficient translation path and the like, the quantized residual DCT coefficients which do not need to be modified are filtered out, useless modification of a carrier is reduced, embedding efficiency of an algorithm is improved, and video quality and algorithm safety are ensured.

The invention relates to a dynamic covalent bond cross-linked oil-displacing polymer and a preparation method thereof. In the present invention, acrylamide is used as the main monomer to increase the length of the molecular chain, thereby increasing the frictional resistance between the molecular chains to improve the viscosity of the solution, and diacetone acrylamide is used as the functional monomer to copolymerize with acrylamide to introduce the ketone carbonyl group into the polymer chain Cross-linking with adipic hydrazide is achieved, resulting in nonionic polymers cross-linked with dynamic hydrazone bonds. The absence of ions in the polymer chain avoids charge shielding and flocculation of various salt ions, and can maintain high viscosity in high salinity water. And the polymer uses the acylhydrazone bond that can be dynamically formed and dissociated as the chemical cross-linking point. As a stimulus-responsive dynamic chemical bond, it not only maintains the reversibility of physical cross-linking, but also has the strength of chemical cross-linking. Reversible formation and breakage can be achieved under external environmental stimuli to overcome the disadvantages of poor shear resistance of traditional polymers.