31results about How to "Improve conversion quality" patented technology

An arrangement for converting speech into text comprises a mobile device (202) and a server entity (208) configured to perform the conversion and additional optional processes in co-operation. The user of the mobile device (202) may locally edit the speech signal prior to or between the execution of the actual speech recognition tasks, by replacing an inarticulate portion of the speech signal with a new version being recording of the portion. Task sharing details can be negotiated dynamically based on a number of parameters.

A device for processing data adapted for being converted between an analog format and a digital format, the device having a scrambling unit adapted for scrambling the data based on at least a part of the data to thereby decorrelate the data in the analog format with respect to the data in the digital format.

The present invention relates to the technical field of geospatial information systems and discloses a method for converting a BIM entity model into a multiple levels of details (LOD) GIS standardized model. Through a method of converting the BIM entity model into the multiple levels of details (LOD) GIS standardized model, the automatic conversion from the BIM entity model to the multiple levels of details (LOD) GIS standardized model is realized, a computer can conveniently execute the geometric semantic information interoperability between two models, the complete conversion and mapping between an IFC format and a CityGML format is realized, and the conversion quality is improved. In addition, the conversion method has the advantages of fast conversion speed and high conversion accuracy, and the practical promotion and application are facilitated.

The editing equipment includes decoding module, audio / video data process module, and encoding module. The decoding module decodes original film to audio data and video data. The audio / video data process module processes the said audio data and video data. The encoding module encodes the processed audio data and video data to object file. The audio / video data process module also includes a unit for filling in or cutting down video data. Through the said unit for filling in or cutting down video data, user carried out relevant filling in or cutting down operations for length and width of video data based on difference value between length width ratio of original film and length width ratio of screen of play back terminal. Using the said unit for filling in or cutting down video data, the invention can eliminate image distortion. Features are: easy of use and high quality of transmission.

The invention discloses a transmitting and rebuilding method of converting a two dimensional (2D) video to a three dimensional (3D) video based on a depth template. The method comprises the steps that a triangle gridding is adopted to describe the depth template and a smoothness template, the smoothness template indicates depth change intensity of each pixel, the triangle gridding description is encoded by an entropy coding and is stored in a picture head of a video sequence, a decoding end regards a 2D picture as a viewpoint, uses the depth template, the smoothness template and the decoded 2D picture to form a depth picture, and a picture of another viewpoint is generated according to the depth picture and the decoded 2D picture. The method of converting the 2D video to the 3D video based on the depth template can effectively help the decoding end to improve quality of converting the 2D video to the 3D video and be low in transmitting cost.

The present invention provides a mechanism to acquire content that has not been converted into print data from a vendor server providing a print service.The mechanism acquires content data corresponding to content identification information from an application server if location information about print data acquired on the basis of the content identification information that is acquired coincides with location information about print data included in notification information that is acquired.

The invention discloses an image domain conversion network and a conversion method based on a generative confrontation network, including a U-shaped generation network, a true-false identification network, and a pair identification network. The image domain conversion process mainly includes the following steps: 1), training U-shaped Generate a network and establish a network model of a U-shaped generating network; 2) input the image to be converted into the network model established in step 1) after normalization processing, and complete the image domain conversion of the image to be converted; the present invention can realize The task of image domain conversion in local areas, and the quality of image local domain conversion is high, the network judgment ability is strong, and the stability of image conversion is strong, which greatly improves the authenticity of generated images.

The embodiment of the present application provides a spoken text processing method, device, server and readable storage medium, by establishing an intermediary object of the written text of the initial spoken text between the original spoken text and the target language translation text, when performing machine translation Machine translation can be carried out based on written texts, so as to take advantage of the high accuracy of machine translation for written texts, improve the translation accuracy and translation efficiency of the initial spoken texts, and reduce the translation cost of manual proofreading. In addition, with the progress of machine translation, the preset written language conversion library can be enriched and updated based on each difference segment between the initial spoken text and the second written language text, which can continuously improve the quality of written language conversion, and then continuously improve The translation quality of the subsequent machine translation further reduces the translation workload of manual correction.

The invention provides a device and a method for quickly converting a YUV420 format into a RGB format. The device comprises a loading module, a separation module, an extension module, a computation module and a combining module, wherein the loading module is used for loading data of YUV420 format into a continuous memory of a graphics processing unit (GPU), the separation module is used for separating the data of YUV420 format, the extension module is used for extending the channel data of a U channel and a V channel, the computation module is used for computing the values of R, G and B channels, and the combining module is used for combining the values of the R, G and B channels into an image of RGB format. The device provided by the invention adopts the GPU to quickly convert the YUV420 format into the RGB format and achieve complete parallel computation, and is applicable to the situations of the YUV420 formats with different resolutions; with the configuration of a relatively low-performance video card, time required for converting the YUV420 format into the RGB format is only within several milliseconds, in the case that the GPU has many resources, multiple paths of YUV420 data can be quickly converted at the same time, the time for converting the YUV420 format into the RGB format is greatly reduced, and the instantaneity is improved under various applications.

The invention provides a track breaking monitoring system of a high-speed railway. The system comprises a MCU for generating a testing signal according to received instruction information and sending to an ultrasonic sending probe, the ultrasonic sending probe for receiving the testing signal and sending an ultrasonic signal to a high-speed railway track according to the testing signal, an ultrasonic receiving probe for receiving the ultrasonic signal from the high-speed railway track and sending the ultrasonic signal to an FPGA processing circuit, and the FPGA processing circuit for performing digital filter and FFT analysis for the ultrasonic signal as well as carrying out frequency spectrum comparison and signal correlation comparing and analyzing for the FFT analysis result so as to determine the track breaking condition of the high-speed railway. According to the track breaking monitoring system of the high-speed railway, the measures of shielding and isolating loops are carried out to increase the signal conversion quality, and the influence of an electromagnetic environment on the detection result is reduced; in addition, the wavelet analysis algorithm is carried out for the FPGA processing circuit, and therefore, the signal receiving sensitivity is high, and the measurement distance is effectively increased.

The present application relates to a lip language recognition method and device based on a double discriminator generation confrontation network, the method includes: deriving face pictures from different angles from the video, and obtaining lip recognition according to different head deflection angles in the face pictures part multi-angle data set; and according to the generator data set, identity discriminator data set, and angle discriminator data set obtained from the lip multi-angle data set, regenerate the confrontation network data model; use the confrontation network data model to be identified Lip recognition is performed on the video, and a 0° lip image is converted; a lip feature vector is extracted from the 0° lip image, and the lip feature vector is modeled and classified to obtain a lip classification result. Partial classification results are exported to recognize the language. The embodiment of the present invention has a visual effect similar to the real environment, can well guide the model to adapt to the actual application environment, and further improve the accuracy of the lip language recognition model.