33 results about "Random text" patented technology

The invention relates to a method for improving rejection capability of a speech recognition system. The method comprises the following steps of collecting various types of noise data; classifying according to the noise types; for different types of noise, respectively training GMMs (Gauss mixed model); assembling various types of GMMs into an integral absorption model; training a statistic language model by various types of relatively random texts, and then establishing a recognition network by WFST (weighted finite state transducer) technique, which is called as an absorption network; connecting the absorption network, the absorption model and an original decoding network in parallel to form a new decoding network; enabling the input original audio frequency to pass endpoint detection and a feature extraction module, so as to generate feature vectors; and competing the feature vectors in the three parts of the decoding network according to an Viterbi algorithm, so as to generate a final recognition result, and effectively reject the noise and an out-of-vocabulary condition. The method has the advantage that on the premise of balancing the recognition efficiency, the effect of rejecting the out-of-vocabulary condition and the invalid input is well realized.

The invention relates to a document encrypting method on the basis of a voice print, which comprises the following steps that: (1) a user remembers a key; (2) the user says a sentence by voice randomly, and a key is extracted from characteristic parameters of the voice print of the voice as an encrypted key; (3) a symmetric encryption algorithm is selected, the key extracted in the step (2) is used to encrypt a document, an intermediate document is generated, a random text string in the step (2) and the intermediate document are packed together to generate a novel document; and (4) the key remembered by the user is used to encrypt the document in the step (3) to form an encrypted document. A voice print document decrypting method comprises the following steps that: (1) the selected symmetric encryption algorithm and the key remembered by the user are used to generate a temporary document; (2) a text string correlative with the encrypted key of the document is separated out from the temporary document and the user says the text string by voice and extracts a key from the characteristic parameters of the voice print of the voice; and (3) the temporary document is decrypted by the key extracted in the step (2) and the selected symmetric encryption algorithm to complete decrypting the document. The document encrypting method and the voice print document decrypting method provided by the invention are convenient to use and have high safety.

A system and method for allowing users to view secure data while excluding others from viewing. The method includes generating a stream of first data frames including secure data content to be viewed on a display device; inserting second data frames within the stream in a manner to thereby render a displayed image of the secure data unreadable on the display device. Shutter glasses, to be worn by an authorized user, are provided to receive signals from the computer device and, in response to received signals, filter out the second data frames thus enabling the viewer to view the secure data to the exclusion of nearby observers. The second frames are generated and interspersed within the first video frames having secure data by software executing on a conventional computing device. The second frames may include random text interspersed according to a regular or irregular pattern, or may include color and intensities to exploit deficiencies of human visual perception.

The invention discloses a random verification code that comprises a random text generation module, a verification code module, a business logic module, a database and the like. According to the invention, user access information is obtained through a system, including user information and access information, such as access address, access time and other dimension parameters. Based on the verification code container selection algorithm, a random text is extracted in the database, verification code characters are then specified in the random text through the verification code module, and a user is notified of verification code selection rules and reminded of the corresponding verification code. If the user is notified of the selection rules, the user only needs to select and input the corresponding verification code to verify. By containing the verification code in the random text, the random verification code improves the security of the verification code system and effectively prevents automatic identification of a machine program.

A cipher method for verifying a decryption of an encrypted user data key used to encrypt user data prior to an encryption of the user data key with an encryption key. The cipher method involves a decryption of the encrypted user data key with a decryption key in response to an initiation of a decryption of the encrypted user data with the user data key as decrypted with the description key, a decryption of the verification text with the user data key as decrypted with the decryption key, and a validation/invalidation of a use of the user data key as decrypted with the decryption key to decrypt the encrypted user data in response to a matched/unmatched comparison of the verification text as decrypted with the user data key and an intermixing of a known text and a random text.

The invention relates to a text line recognition method and device, a readable storage medium and electronic equipment. The text line recognition method comprises the steps that at least one to-be-detected text line image is input into a preset neural network model, a detection result output by the preset neural network model is obtained, and the detection result is a character string in the to-be-detected text line image, wherein the preset neural network model is obtained by training by taking the synthesized image sample as a training sample. A composite image sample comprises a sample, a positive sample, a difficult negative sample and a common negative sample. According to the technical scheme, a synthetic image sample comprising a sample, a positive sample, a difficult negative sample and a common negative sample is used as a training sample of a preset neural network, the training sample of the preset neural network is enabled to be the associated image data, so that more supervision information is introduced, and the training effect of the preset neural network model is improved in comparison with the mode of adopting mutually independent random text line images as the training data for training.

The invention relates to an image identifying code achievement method based on text adhesion and visual compensation.The image identifying code achievement method comprises the steps of identifying code and candidate code generation, identifying code input and identifying code identification.The step of identifying code and candidate code generation comprises the sub-steps that a server side generates candidate characters and correspondingly generates serial numbers S0; identifying characters are selected form the candidate characters, and the serial numbers S0 are correspondingly set; the identifying characters are randomly arranged, and the output positions of the identifying characters are set; the character sizes, character fonts, colors, shapes and image backgrounds and interference backgrounds of the identifying characters are set; image identifying codes for random text adhesion are formed by combining with the identifying characters, the image backgrounds and the interference backgrounds, each candidate code correspondingly generates an independent character image, and the character images are directly output to a web page to be identified and are combined with the identifying codes to serve as web page candidate codes.By the adoption of the image identifying code achievement method, the problem that identifying codes are not easily identified and cracked by machines automatically but are easily identified and input by users is solved.

The present invention discloses a secure login method. The method comprises: obtaining first voice data generated by a user by reading a preset feature text, and determining a voiceprint feature of the first voice data; obtaining second voice data generated by a user by reading a random text, and determining the voiceprint feature of the second voice data; and determining the matching degree of the voiceprint feature of the first voice data with the voiceprint feature of the second voice data, and when he matching degree is greater than the preset matching degree threshold, allowing the user to login. The present invention also discloses a secure login apparatus.

A HRPD network access authentication method based on CAVE algorithm is proposed. An AT generates a random number “RAND” necessary for the calculation of AUTH1 by using the “Random text” included in the CHAP Challenge message. A UIM card works out the AUTH1 by using the random number “RAND” and the SSD_A in the UIM card. The AT carries the AUTH1 by the Result field of the CHAP Response message. The AN-AAA generates the random number “RAND” necessary for the calculation of AUTH2 by using the “Random text” included in the Radius Access Request message. The AN-AAA works out the AUTH2 by using the random number “RAND”. The two results AUTH1 and AUTH2 is compared. If they are just the same, the authentication on the AT passes. Otherwise, the AT is rejected to access.

Disclosed are a speech recognition verification device and a speech recognition verification method, which verify speech recognition results by executing artificial intelligence (AI) algorithms and/or machine learning algorithms in a 5G environment connected for Internet-of-Things. According to an embodiment, the speech recognition verification method includes converting a verification target text item to a verification target spoken utterance by applying a preset utterance condition, analyzing the verification target spoken utterance and outputting a recognition result text item corresponding to an analysis result, and verifying speech recognition performance through comparison between the verification target text item and the recognition result text item. According to the present disclosure, the speech recognition result may be verified objectively by using a spoken utterance generated with random text and various utterance conditions as input of speech recognition.

The present invention discloses a text-to-speech system that provides output variability. The system can include a finite state grammar, a variability engine and a text-to-speech engine. The finite state grammar can contain a phrase role consisting of one or more phrase elements. The phrase rule can deterministically generate a variable text phrase based upon at least one random number. The phrase rule can include a definition for each of the phrase elements. Each definition can be associated with at least one defined text string. The variability engine can construct a random text phrase responsive to receiving an action command, wherein said finite state grammar is used to create the text phrase. The variability engine can also rely on user-specified weights to adjust the output probabilities. The speech-to-text engine can convert the text phrase generated by the variability engine into speech output.

The invention discloses an OCR training sample generation method, device and system, and relates to the field of computer vision. The method comprises the following steps: a character contour extraction step: extracting all character contours based on an original image, determining an erasure area mask by combining an erasure area coordinate, and obtaining a repair area mask; an image repairing and filling step: performing image repairing and filling according to the mask of the repairing area and the pixel information around the repairing area to obtain a background template after the characters are erased; and a random text generation step: generating a random text in each generation area so as to obtain a new sample picture and a labeling information file corresponding to the new sample picture. According to the method, a character contour extraction algorithm, an image restoration technology and other technologies are combined, background information of an original picture is fully utilized, a high-quality training picture is generated, meanwhile, an annotation file (including character content and position information) corresponding to the picture is generated, the burdensome and labor-consuming annotation work is omitted, and the method can be directly used for OCR model training.

The invention relates to the field of communication, and provides a verification method. The verification method comprises the following steps: receiving, by a verification code server, a verificationcode acquisition request sent by a client; generating and storing a system verification code, wherein the system verification code is determined by a string of random texts; encoding the system verification code into n original pictures to generate k verification code pictures, wherein k and n are integers greater than 2, and k is less than or equal to n; sending the k verification code picturesto the client; receiving a verification request sent by the client, wherein the verification request carries a verification code input by a user according to the k verification code pictures; and determining whether the verification code input by the user is correct according to the system verification code. By adoption of the above scheme, the randomly generated system verification code is codedinto the n original pictures, which not only increases the difficulty of machine recognition, but also can ensure the security of the verification code.

An apparatus and a method for validating compressed encrypted archive keys is described. In one embodiment, a pseudo-stream is generated for an archive. The pseudo-stream is made of a small amount of random text. The pseudo-stream is attached to a stream of the archive. The pseudo-stream and stream are compressed using the compression algorithm that includes validation data. The compressed pseudo-stream is then enciphered with an archive key.

The invention discloses a carrier-free text steganography method and a carrier-free text steganography device for a component histogram. The carrier-free text steganography method comprises the following steps: 1, constructing a codebook of a text library; 2, preprocessing the secret information; 3, providing a hiding method based on random text selection; 4, performing information hiding; 5, constructing a secret index; and 6, extracting secret information. Compared with an existing text carrier-free information hiding technology, the invention has the advantages that all secret information can be hidden in one text at most, the hiding capacity is greatly improved, and the hiding success rate is increased by using a mixed index structure instead of a natural language processing method in the extraction process.