80 results about "Perceptual hashing" patented technology

The invention discloses an intelligent monitoring system and method for nursing infants, and the method comprises the following steps: S1, collecting continuous frames of infrared images, carrying out the preprocessing, and extracting feature vectors; s2, constructing a classification model for identifying infant micro-motion types, and training the classification model to ensure accurate identification of the classification model; s3, acquiring a real-time infrared monitoring video image of the infant, performing similarity analysis on continuous frames, judging the range of the continuous frames of the micro-action, and extracting continuous frame images; and S4, feature vectors are extracted from the extracted continuous frame images, the type of the infant micro-motion is identified, and user side reminding and analysis are carried out. Aiming at low-light environment and micro-motion detection in baby nursing, through combination of infrared imaging and optical flow feature extraction and fusion of a perceptual hash array and a Hamming distance, real-time and reliable baby state monitoring service is provided for parents or caregivers, and the safety and convenience of baby nursing are enhanced.

The invention provides a distributed multi-source digital content tamper-proof Hash evidence storage method based on a block chain, and relates to the technical field of content tamper-proof, and the method comprises the steps: standardizing multi-source digital content, constructing an evolution pedigree tree, implanting a detector to capture change features, and tracking a propagation path; constructing a dynamic perceptual hash network to generate a dynamic hash value containing associated information; and verifying and generating an evidence storage record through a progressive consensus mechanism. According to the invention, efficient tamper-proof evidence storage of the multi-source digital content is realized, and the data traceability and verification efficiency are improved.

This disclosure relates to approaches for managing generative artificial intelligence content. Some aspects relate to managing and enforcing licensing conditions for various actors in generative artificial intelligence, such as content creators, model owners, and platform providers. Some aspects relate to lineage tracing, in which an asset is analyzed to determine what other assets it was based on and / or to determine which other assets were generated based on a given asset. Lineage tracing can be based on one or more of metadata analysis, hash analysis (such as perception hash analysis), or vector similarity. In some implementations, a vector embedding model is used to generate vector embeddings of assets, and the vector embeddings are compared to identify similar assets.

The invention relates to the technical field of model watermarking, and discloses a multi-modal model watermarking method based on synonym replacement, which comprises the following steps of: inputting a given image and a prompt into a multi-modal model to generate a text sequence for describing the image or a text sequence for answering a question proposed based on the image; watermark information can be detected in the text sequence; the method comprises the following steps of: mapping a secret image into a hash value as a private key by using a perceptual hash algorithm; dividing the vocabulary by using a private key and changing an output sample to embed watermark information; and checking whether the watermark exists or not by counting the number of words in a green list in a text sequence output by the multi-modal model. According to the method, the watermark can be embedded into the multi-modal model only by changing the text output sample, retraining is not needed, and detection can be carried out without accessing the model, so that the method is suitable for all mainstream multi-modal model structures, and strong generalization is provided.

The invention discloses an edge AIGC environment-oriented verifiable ownership storage method and framework. The method comprises the following steps: constructing a perceptual hash and verifiable storage ownership framework comprising a user, an edge AIGC generation server, an ownership evidence calculation node and a distributed evidence warehouse; a user submits an AIGC generation request to the edge AIGC generation server, and the edge AIGC generation server generates a target AIGC image; performing perceptual feature extraction on the target AIGC image to generate a unique hash sequence; binding the unique hash sequence with a user ID and sending the unique hash sequence and the user ID to the distributed evidence warehouse, and after the user confirms that the target AIGC image is received, storing binding information of the unique hash sequence and the user ID by the distributed evidence warehouse; and when a suspected infringement image appears, calculating a hash sequence of the suspected infringement image and querying an originally stored unique hash sequence to judge and determine whether infringement occurs, so as to realize verifiable storage of the ownership of the AIGC image.

The invention relates to the field of hybrid coding of video engines, and provides a hybrid coding processing method, system and equipment based on a video engine, which comprises the step of intelligently switching an H.265 inter-frame predictive coding mode and an MJPEG (Multijoint Joint Photographic Experts Group) intra-frame compression coding mode by dynamically monitoring the proportion of a motion area in a video picture. When a scene is static, discrete cosine transform is adopted to compress space redundancy, motion vector compensation is started to eliminate time redundancy when motion is violent, and meanwhile, key frames are doubly screened by using pixel difference analysis and a perceptual hash algorithm, and repeated frames of visual redundancy are eliminated. And finally, space-time association is established for the optimized double-code-stream data through a timestamp index system, and efficient storage and accurate reconstruction of the mixed code stream are realized. The problems that the coding efficiency is low, no self-adaptive coding mode exists, redundant frames are not optimized, and code stream management is insufficient are solved.

The invention relates to the technical field of block chains, and discloses an artwork chaining evidence storage and source tracing method, which comprises the following steps: an artwork chaining stage: a seller signs an artwork needing to be chained, and broadcasts a signature and a perceptual hash value of the artwork to a block chain; the committee verifies whether the artwork is repeatedly linked, if not, the artwork is formally linked, and the artwork is sold on a transaction platform specified by a seller; in the artwork transaction stage, the purchaser places an order to purchase artworks; the transaction platform issues the transaction information to the block chain; the seller delivers goods; the purchaser verifies the received artwork, and if verification succeeds, the transaction is completed; in the artwork traceability stage, an inquirer inquires historical records of artworks through a transaction platform or a block chain, and the historical records comprise initial chaining time and previous transaction information. According to the invention, right confirmation, source tracing and trust of artwork transaction are realized based on the block chain.

The invention discloses a medical image robust watermarking algorithm based on a re-parameterized high-efficiency attention network, and relates to the field of image data processing methods, and the algorithm comprises the steps: obtaining an original medical image; a pre-trained RepEA-Net model is used for extracting multi-scale deep features with good generalization performance from an original medical image, and then a mean perceptual hash algorithm is used for generating feature vectors. Obtaining an original watermark image; the watermark image is encrypted by using a Logistic chaotic encryption algorithm, the security of the watermark image is improved, and finally, the encrypted watermark image and the binary feature vector are subjected to XOR operation to generate the zero watermark. Through the innovative design of the re-parameterization efficient attention network, the contradiction among robustness, generalization and image integrity of a traditional medical image watermarking algorithm is broken through, and a solution with technical advancement and clinical practicability is provided for copyright protection and safe transmission of medical images in the medical Internet of Things.

The invention discloses a pilot protection method, system and device for a new energy station sending-out line, and relates to the technical field of new energy station sending-out line protection.The method comprises the steps that transient fault current signals of the new energy station side and the system side on the two sides of a fault are decomposed into different frequency components through DCT; obtaining a two-dimensional DCT coefficient corresponding to the current data at each time point, and generating a binary image; the binary images are flattened into one-dimensional vectors to obtain respective perceptual hash values, the Hamming distance between the two perceptual hash values is calculated and compared to judge the similarity of waveforms, and pilot protection is carried out according to the similarity; according to the method, fault positioning of a sending-out line and accurate action of a protection device are achieved, the method is suitable for a network-forming type large-scale new energy sending-out system, and rapid and efficient fault processing is achieved.

The invention relates to the technical field of information retrieval, in particular to an image retrieval method and device based on depth difficulty perceptual hash and a storage medium. In the method, in order to comprehensively model a global geometric structure of an embedded space, self-adaptive difficulty perception enhancement is carried out on an original data sample, and the difficulty level is dynamically adjusted by utilizing linear interpolation for embedding; difficulty perception feature synthesis is carried out to generate enhanced data samples which are consistent in semantics and reserved in labels, and loop training is carried out on the enhanced data samples; the strategy not only makes full use of potential information in all training samples, but also systematically challenges learned metrics with the difficulty of adaptive calibration, ensures continuous optimization on the whole feature manifold, and promotes generation of similarity-preserving hash codes.

A non-contact body weight measuring method based on two-dimensional pig contour image information comprises the following steps: a, data acquisition: acquiring a two-dimensional image of the back of a pig by using a camera mounted above a pigsty, performing stream extraction on the camera through an RTSP protocol, filtering similar frames in a video stream by using a perceptual hash technology, and obtaining a two-dimensional image of the back of the pig; obtaining a pig image and visual mark information of the back of the pig; acquiring pig numbers by using a radio frequency identification technology and recording the pig numbers in a database; obtaining the serial number of each pig in the pig image in combination with the pig back visual mark information; b, contour extraction: acquiring back contour information of the pig in the image by using a contour extraction network; and c, measuring the weight, namely predicting the weight of the pig by using a weight estimation network based on the fine contour information. According to the method, the weight of the pig in the actual production environment is measured in real time under the conditions that the activity of the pig is not restrained and the pig is not required to keep a specific posture, no stimulation is caused to the pig, and the method is efficient and accurate.

The invention discloses a video content protection method and device based on a multi-channel image feature security mark, and relates to the technical field of video content security management, and the method comprises the steps: obtaining an original video, and checking the original video; acquiring a video frame image of the audited video, and performing fusion, dimension reduction, compression and encryption processing on multichannel high-order features of the video frame image through a perceptual hash algorithm, an interpolation hash algorithm, a compression coding principle and a national cipher SM4 algorithm to generate a dynamic security mark; embedding the dynamic security mark into the video frame image to generate an approved video; and extracting the dynamic security mark of the currently played video, and checking the consistency between the feature information extracted from the dynamic security mark and the feature information of the played video to judge whether the video is compliant or not. The video content is protected, the security of the video is improved, the compliance check of the video content is realized, and the accuracy and efficiency of compliance judgment of the video content are improved.

A system for the tamper-proof verification of digital media, comprising the following functions: a) generating a cryptographic pattern by a physical device, wherein the pattern is derived from a timestamp, a device identification, and a random value, and signed using a cryptographic signature procedure with a non-extractable key stored in a security element; b) transmitting the pattern by the physical device using a physical signal that is not perceptible or only partially perceptible to humans; c) capturing the pattern by a digital capture device during the recording of a digital medium; d) recognizing and extracting the pattern from the recorded digital medium by an application; e) calculating at least one cryptographic hash and at least one perceptual hash of the digital medium; f) anchoring the verification data in a tamper-proof data storage device;g) Verification of the digital medium by comparing the calculated hashes with the anchored values ​​and checking the match of the extracted pattern with the generated pattern.

A system and method of responding to fuzzing including receiving a fuzzing input while in fuzzed mode, performing a perceptual hashing function using the fuzzing input to generate a perceptual hash value, selecting an action from a list of actions using the perceptual hash value, and performing the selected action in response to the fuzzing input. Parameters may be generated using the perceptual hash value for actions that use parameters. Instead of normal hashing, perceptual hashing generates the same hash value for substantially similar fuzzing inputs so that corresponding fuzzing response actions appear to be random but instead are intentional. Hardware or software version numbers may be combined with a shared secret key and hashed using a non-perceptual hashing function to further impede comparison analysis by a fuzzer. Some embodiments combine perceptual hashing with non-perceptual hashing, such as cryptographic hashing or the like.

The invention belongs to the technical field of copyright protection, discloses a lightweight active defense picture copyright protection method, and aims to solve the problems of penalty after infringement, huge storage overhead caused by full-node storage and undifferentiated copyright and use right in copyright protection. The invention provides a lightweight active defense picture copyright protection scheme. On one hand, the scheme is based on the perceptual hash technology, active defense is achieved, and malicious resale behaviors are completely eradicated before infringement behaviors occur; on the other hand, storage optimization is realized based on a chameleon hash algorithm, and transaction information of the same copyright is added to an original block under the condition that a block head is not changed; meanwhile, a copyright chain and a use right chain are designed, the copyright can be transacted, and the use right is only allowed to be downloaded and checked after identity verification.

The invention provides a multi-dimensional construction data storage system and method, relates to the technical field of data storage, and realizes significant improvement of construction data storage efficiency and breakthrough of intelligent management capability by introducing multi-dimensional feature extraction, resolution factor calculation and a dimension perceptual hash algorithm. Firstly, original feature values of four dimensions are extracted from each construction data unit, standardized coding processing is carried out, a unified dimension code is generated, the precision degree and the association strength of data are quantitatively evaluated by calculating a time resolution factor, a space resolution factor and a coding similarity factor, and therefore a storage strategy is dynamically adjusted, and the storage efficiency is improved. According to the dimension perceptual hash algorithm, logic storage coordinates are intelligently generated through the factors, data with similar characteristics are naturally gathered on physical storage, data items are temporarily stored by maintaining an ordered structure on the memory management level, sorting is conducted according to the logic coordinates, disk random write-in operation is effectively reduced, and write-in performance is improved.

Methods and systems for efficiently generating and verifying 3D perceptual hashes of 3D models are provided. Embodiments include providing a 3D model of an object, the 3D model including a plurality of mesh vertices; finding a minimal bounding sphere substantially enclosing the plurality of mesh vertices and having a center corresponding to the center of mass of the object; calculating, based on the minimal bounding sphere, one or more sub-spheres concentric with the minimal bounding sphere and evenly distributed inside the minimal bounding sphere. Further, for each sphere: projecting one or more normal vectors from the center of the sphere to one or more object faces intersecting the sphere; calculating, based on the one or more normal vectors, one or more first distances from the center of the sphere to the one or more object faces; extending the one or more normal vectors up to a radius of the sphere to project the one or more normal vectors as one or more dome points on the sphere; calculating one or more second distances between the one or more dome points; and then mapping the one or more first and second distances corresponding to all spheres to a hash space to generate a 3D perceptual hash value of the 3D model.

The application relates to a local geometric fine correction method suitable for high-resolution remote sensing images, relates to the field of remote sensing image geometric fine correction, and solves the problems that the existing method is calculated for the whole image, time and calculation resources are wasted, error matching is prone to occurring by using the existing matching method, geometric fine correction fails and the like, the rational polynomial RPC coefficient of the whole image is used for calculating the longitude and latitude range of the whole image; the longitude and latitude of a local area interested by a user are used for generating an image of the local area with rough geographic coordinates by using an indirect resampling method; a perceptual hashing algorithm is used for estimating the rough coordinate offset between the local area image and a reference base map, the part of the image corresponding to the local area image is cut from the reference base map, and the local area image to be processed is registered with the local reference base map by using Fourier transform blocking. The method has significant improvement in speed, accuracy and stability, and is suitable for various meter-level and sub-meter-level high-resolution remote sensing images.

The invention discloses a secrecy inspection method and system based on an OCR (Optical Character Recognition) edge server, and belongs to the technical field of computer and information security. The method comprises the steps of obtaining a target image, calculating a perceptual hash value of the target image, and performing preliminary screening on the target image through a built-in lightweight OCR model; and generating sensitive file alarm information of the target image based on a preliminary screening result, or uploading the target image and the perceptual hash value thereof to an edge server for processing, performing duplicate checking identification on the target image by the edge server based on the perceptual hash value, and if a duplicate checking identification result is that historical repeated data exists, sending the sensitive file alarm information to the target image. If yes, a perceptual hash value corresponding to the historical repeated data is used as an identification result to be returned to the terminal equipment, otherwise, a precise OCR model is called to carry out full-text identification on the target image, and the identification result is returned to the terminal equipment; and the terminal device generates sensitive file alarm information of the target image in response to the identification result. According to the invention, the throughput and response speed of the system can be effectively improved on the premise of ensuring the security of confidentiality check.

The invention discloses a picture interception method and device, equipment and a storage medium, and belongs to the field of application software. The method comprises the following steps: calculating perceptual hash of a to-be-uploaded picture, and obtaining hash fragments of the perceptual hash; according to the Hash fragment of the to-be-uploaded picture, matching from the at least two Hash buckets to obtain a target Hash bucket; the hash bucket corresponds to a bucket hash value, the hash bucket is used for storing a reference picture, and a hash segment of the reference picture is the same as the bucket hash value of the hash bucket; the target bucket hash value of the target hash bucket is the same as the hash segment of the to-be-uploaded picture; under the condition that a reference picture similar to the to-be-uploaded picture exists in the target hash bucket, carrying out steganography content analysis on the to-be-uploaded picture; and intercepting the to-be-uploaded picture under the condition that the steganography content exists in the to-be-uploaded picture. According to the method, the picture with the steganographic content can be intercepted quickly and efficiently.

The invention discloses a similar image search method, system and device based on perceptual hash and a medium thereof, and the method comprises the steps: constructing a multi-scale space for an input initial image, generating a Gaussian pyramid, and obtaining a Gaussian difference pyramid based on the Gaussian pyramid; screening stable key points in the Gaussian difference pyramid; according to the distribution density of the screened stable key points, delimiting a densest area in the initial image, and intercepting a minimum bounding rectangle as a target screenshot image; extracting multi-modal features of the initial image, and performing feature splicing to generate a joint feature vector; preprocessing the initial image and the target screenshot image; generating a corresponding Hash character string based on the preprocessed image and the joint feature vector; and comparing the Hash character string of the image to be retrieved with the Hash set in the database, and counting the image of which the Hamming distance is smaller than or equal to a preset threshold value as a similar result to be output. According to the method, the search accuracy of the screenshot image is improved, so that complete original image information is obtained.

The application provides a robust watermark and password signature joint embedding and proving method and device for AIGC images, and the method comprises the following steps: constructing a structured watermark load and performing attack-resistant redundant coding to obtain a redundant load sequence, and calculating a first hash value of the structured watermark load; embedding the redundant load sequence into a hidden space domain feature of an AIGC generation model at at least one preset embedding stage of intermediate features of the AIGC generation model to generate an initial AIGC image carrying a hidden space domain watermark; embedding a watermark synchronization code, load backup information and a signature record identifier in medium and low frequency coefficients of frequency domain coefficients obtained after frequency domain transformation of the initial AIGC image to generate an intermediate AIGC image carrying a cross-domain double-layer collaborative robust watermark; and binding a password signature, a reference perceptual hash value and the signature record identifier to a file structure of the intermediate AIGC image and / or an external notarization record to generate a target AIGC image carrying the cross-domain double-layer collaborative robust watermark and the password signature.

According to the image-text multi-mode data downloading duplicate removal method and device, before image data are downloaded, downloaded image links are filtered out, and downloading of duplicate data is reduced. After the image data is downloaded, the image perception hash value is calculated, the repeated image data is filtered out, and then the image tags with the same image link or the image perception hash value are merged into the image tag list, so that the data storage space is reduced, the tag information of the image is enriched, and the image tag list is enriched. The defects of storage resource waste, low downloading efficiency, low retrieval efficiency and the like caused by general repeated storage of image-text multi-modal data in a network environment in the prior art are overcome, and a good application effect is achieved.

The invention relates to the technical field of image processing and computer vision, in particular to a general image comparison method and device combining high-level semantics and low-level features. The method comprises the following steps: acquiring an image object to be compared, and setting a similarity threshold value; a pre-trained ResNet model is adopted to extract high-level semantic features of image objects, cosine similarity is calculated, and candidate image pairs are obtained through preliminary screening; respectively processing and comparing the candidate image pairs by utilizing perceptual hash, average hash, mean square error and structural similarity comparison algorithms, and screening potential image pairs again; according to a preset weight distribution scheme, performing weighted average on the plurality of similarities obtained by the plurality of algorithms of the potential image pair to obtain a final similarity; finally, the images corresponding to the similarity greater than the similarity threshold are similar image pairs. The method effectively solves the limitation of contrast, brightness, rotation, zooming and the like in the prior art, and can be widely applied to scenes such as document similarity detection, copyright protection, content auditing and the like.

The present invention relates to the field of information retrieval technology, and more particularly to an image retrieval method, device, and storage medium based on deep difficulty-aware hashing. In the present invention, adaptive difficulty-aware enhancement is performed on original data samples to comprehensively model the global geometric structure of the embedding space, dynamically adjusting the difficulty level by linear interpolation of the embedding. Difficulty-aware feature synthesis is also performed to generate semantically consistent and label-preserving enhanced data samples for cyclic training. This strategy not only fully utilizes the potential information in all training samples, but also systematically challenges the learned metric with an adaptively calibrated difficulty, ensuring continuous optimization across the entire feature manifold and promoting the generation of similarity-preserving hash codes.

The application discloses a similar image search method, system and device based on perceptual hashing and a medium thereof. The method comprises constructing a multi-scale space for an input initial image and generating a Gaussian pyramid based on the multi-scale space, obtaining a Gaussian difference pyramid based on the Gaussian pyramid, screening stable key points in the Gaussian difference pyramid, delimiting a densest region in the initial image and intercepting a minimum circumscribed rectangle as a target screenshot image according to the distribution density of the screened stable key points, extracting multi-modal features of the initial image, performing feature splicing to generate a joint feature vector, preprocessing the initial image and the target screenshot image, generating corresponding hash strings based on the preprocessed images and the joint feature vector, comparing hash strings of images to be searched with a hash set in a database, and outputting images with a hamming distance less than or equal to a preset threshold as similar results. The application improves the search accuracy of screenshot images to obtain complete original image information.

The present invention provides a method, a storage medium, and a computing device for determining similar dark web products based on multi-modal fusion features. The method includes: Step 10, collecting dark web product data and classifying the products; among them, the collected dark web product data includes product text and product pictures, and an md5 value is generated for the collected product pictures; Step 20, based on the dark web product data collected in Step 1 and the product classification results, calculating the perceptual hash fingerprint value of the product pictures and the Word2Vec sentence vector value of the product text; Step 30, calculating the product similarity based on the md5 value of the product pictures, the perceptual hash fingerprint value of the product pictures, and the Word2Vec sentence vector value of the product text. By constructing a similarity calculation method that fuses product pictures and product texts, the present invention can solve the problem that it is difficult to determine similar products due to the blurred product pictures and simple text information on the dark web.

The invention discloses a power transformation operation and maintenance intelligent management system and method, and relates to the technical field of intelligent management. A thermal imaging image of first equipment is collected and preprocessed, and a first image is obtained and converted into a first image matrix; performing feature extraction on the first image through a perceptual hash algorithm to obtain corresponding temperature features; acquiring a second image by acquiring a second device, and performing feature extraction to obtain a temperature feature of the second image; calculating the Hamming distance between the second image and all the first images in the standard image library, obtaining the minimum Hamming distance, comparing the minimum Hamming distance with a preset similarity threshold value, obtaining the operation state of the second equipment, and outputting a recognition result; formulating an equipment maintenance task for the identified abnormal equipment, calculating the matching degree between each maintainer and each equipment maintenance task, and selecting the maintainer corresponding to the maximum matching degree to execute the equipment maintenance task; and constructing a user management interface, and displaying the equipment operation state and the equipment maintenance task progress.

The invention relates to an image data intelligent enhancement method and system based on a large model, and the method comprises the following steps: S1, obtaining power transmission line hardware fitting image data collected based on an unmanned plane, carrying out the image segmentation through employing a Mask R-CNN, and removing invalid images; s2, identifying and repairing abnormal images with overexposure and underexposure by analyzing brightness histogram distribution characteristics of the images; s3, adopting a bilinear interpolation algorithm to adjust all images in the brightness correction image set into a unified target size, and adopting a perceptual hash algorithm to combine with cosine similarity calculation to identify and remove repeated images to obtain a standardized de-duplicated image set; s4, carrying out ControlNet-based conditional image generation enhancement on the standardized de-duplicated image to obtain an enhanced image set; and S5, for the enhanced image set, using a pre-training detection model to generate an initial annotation, and combining manual calibration to obtain a complete annotation data set. According to the invention, the hardware fitting damage detection precision and efficiency are effectively improved.

The invention belongs to the technical field of medical image transmission, and particularly discloses a zero watermark method and system based on an improved lightweight neural network MobileNetV3, and the method comprises the following steps: S1, collecting an original medical image; s2, encrypting the original watermark image by using a Logistic logic chaotic system; s3, inputting the original medical image into an improved lightweight neural network MobileNetV3, and obtaining multi-scale depth features of the medical image; s4, based on the multi-scale depth features of the medical image, creating corresponding binary feature vectors by using a perceptual hash algorithm; and S5, performing XOR operation on the encrypted watermark image by using the binary feature vector to generate a zero watermark. By adopting the technical scheme, based on the improved lightweight neural network MobileNetV3, the method can focus on images of channel and spatial dimensions, the generalization ability and performance of the method are enhanced, it is ensured that the extracted medical features have excellent robustness, and the method is suitable for being applied to mobile medical care.