52results about How to "In line with subjective feelings" patented technology

The invention provides a quality estimation method of exit-entry digital portrait photos. The quality estimation method comprises the following steps: the format size and file size detection of each digital portrait photo, face detection, eye positioning, face horizontal centering detection, the exposure detection of global luminance deviation and partial luminance deviation, the clearness detection of contrast evaluation and ambiguity evaluation and the color detection and the background detection for detecting the skin naturalness in an Lab color space and judging whether colors in an image face area are qualified or not. The quality estimation method has the advantages that a quality estimation module of the digital portrait photos, specific quality detection items, detection flows, estimation indexes and estimation methods are established, which is different from the prior quality estimation method of the digital portrait photos relying on subjective estimation. The invention judges the quality of each digital portrait photo without relying on a standard image, obtains the judging result in accordance with the subjective feeling of people and the requirements of a face recognition system, thereby favorably satisfying the business requirements for managing the exit-entry digital portrait photos.

The invention discloses an image coding method and system based on dictionary learning. The method comprises the steps that image blocks are randomly selected from an image sample to carry out sparse representation based on batch processing, and a dictionary matrix is solved through an alternative optimization method; according to the obtained dictionary matrix, the sparse representation processed in batch is utilized to carry out sparse representation on the image blocks to be coded, and quantification entropy coding is carried out on nonzero coefficients in sparse coefficient vectors; decoding and sparse reconstruction are carried out on the coded image at a decoding end according to the obtained dictionary matrix. The system comprises a dictionary learning module of sparse representation based on batch processing, an image coding module and an image decoding module. According to the method and system, the problem that the training sample in dictionary learning is large in scale is resolved, meanwhile reconstruction errors of the image sample are reduced, and rate-distortion performance of image compression is significantly promoted.

The invention discloses a no-reference objective video quality assessment method. According to the assessment method, video files are processed and calculated respectively based on a plurality of quality parameters for assessing video quality, and the values of the calculated quality parameters are synthesized into a final video quality assessment value, wherein the quality parameters comprise block effect, signal-to-noise ratio, contrast and edge energy. The method comprises the following steps of: performing sub-frame pre-processing on the video files, then extracting the values of the quality parameters, and finally acquiring a numerical value capable of objectively assessing the video quality by linear fitting. The quality of the distorted video is assessed according to a human visionsystem and signal-to-noise ratio, block effect, contrast and edge energy and the like; and the method is higher in accuracy, more comprehensive in analysis and more consistent with the subjective feeling of the human being, and the effect of the method can be matched with full reference assessment.

The invention provides a method and a device for detecting defects of a display screen based on human visual characteristics. The method acquires an image of a display screen to be detected. Frequencydomain filter kernel is created by using human eye contrast sensitivity function. Fourier transform is performed on that screen image to be detected to obtain the transformed screen image to be detected; The frequency domain filter kernel is used to filter the transformed display screen image to obtain the filtered display screen image. The frequency domain filter kernel is used to filter the transformed display screen image to obtain the filtered display screen image. Inverse Fourier transform is performed on the filtered screen display image to obtain a target display screen image; According to the image of the target display screen, the segmentation threshold is calculated according to the preset algorithm. According to the segmentation threshold, segmenting the target display screen image to obtain a plurality of display screen sub-images; All the sub-images of the display screen are screened to obtain the defect image of the display screen. Using the human eye contrast sensitivity function, a frequency domain filter kernel is created to make the defect image accord with the human eye visual characteristics and improve the accuracy of the detection results.

The invention discloses a method and device for assessing multimedia quality. The method comprises the step of obtaining benchmark quality and ultimate quality of a video of a multimedia sequence and benchmark quality and ultimate quality of of an audio of the multimedia sequence, the step of determining the benchmark quality of the multimedia sequence based on the benchmark quality of the video and the benchmark quality of the audio, the step of determining a distortion value of the multimedia sequence according to the benchmark quality of the video, the ultimate quality of the video, the benchmark quality of the audio and the ultimate quality of the audio, and the step of determining the multimedia quality of the multimedia sequence according to the benchmark quality of the multimedia sequence and the distortion value of the multimedia sequence. The method and device for assessing the multimedia quality can directly reflect the distortion of the multimedia sequence, conforms to the subjective feeling of people, and then can accurately and effectively assess the multimedia quality.

An embodiment of the invention discloses a method and apparatus for obtaining video coding compression quality. The method may include obtaining video stream information including a video frame type, a video frame size, a frame rate and a bit rate; according to the video stream information, calculating video content complexity including time complexity or space complexity or time complexity and space complexity; and calculating video coding compression quality according to the bit rate, the frame rate and the video content complexity. In the embodiment of the invention, the video coding compression quality is obtained requiring only the video frame information, the bit rate, the frame rate and the video content complexity of a video stream, the video coding compression quality is available for use in subsequent video quality assessment, the complexity of the video quality assessment is greatly reduced and real-time assessment can be achieved. Furthermore, with the influence of video content characteristics (i.e. video content complexity) and the frame rate being considered, the assessed coding compression quality accords better with subjective visual feelings of humans.

The invention discloses a non-reference image quality evaluation method based on a convolutional self-coding network. The concrete steps of the invention are as follows: constructing a convolutional self-coding network; constructing a fully connected neural network; generating a pre-training set, a training set and a test set; training a convolutional self-coding network and fully connected neuralnetwork; and making row quality evaluation of distorted images in the test dataset. The convolutional self-encoding network is used to encode non-reference images and image blocks thereof respectively, global semantic features and local distortion features are extracted from the coding of non-reference images and their blocks by using the fully connected neural network, and the two features are fused, and the fused features are mapped to perceptual quality scores by using the fully connected neural network. The evaluation results are more consistent with the subjective feelings of people.

The invention discloses a blocking effect analysis method. The method comprises the following steps: inputting a decoded image; calculating brightness differences on boundaries among pixel blocks; utilizing a blocking effect visibility function to process the brightness differences on the boundaries among the pixel blocks; utilizing a Weber-Fick inner law to calculate the subjective sensation of a human vision system to visible blocking effects; and integrating the blocking effects on all the boundaries in the image to obtain the blocking effect evaluation analysis results of the whole image. The blocking effect analysis method and the system can accurately realize the blocking effect evaluation so that results can more suit the subjective sensation of the eyes.

The invention discloses a loudspeaker abnormal sound detection method and system and mainly aims at solving the problem that a result obtained by use of a traditional loudspeaker abnormal sound detection method is inconsistent with the judgment result of the subjective sensation of human eyes. The invention relates to a loudspeaker abnormal sound detection method and system based on a psychoacoustic sharpness parameter; according to the method and the system, the abnormal sound of the loudspeaker is detected by calculating the high-order sharpness distortion of the loudspeaker. The loudspeaker abnormal sound judgment result obtained by use of a high-order sharpness distortion algorithm in the detection method is more consistent with the subjective sensation of the human eyes, and therefore, the misjudgment ratio can be reduced and the correctness of the judgment result can be improved.

The invention provides a combined big screen front-direction projecting display method and control device thereof. The method comprises: placing the projector, making the maximal projecting area of the projector partly overlapped; generating the testing sub images and projecting the images to the display plane via the projectors; manually adjusting various sub-screen which can be correctly spliced into a complete picture of the big screen; to manually adjust the screen after the test with the original sub-sub-screen comparison of the test calculation, the various projector display deformation parameters, and save; in accordance with the correction process constitute a test sub-screen picture of the whole approach will be required to show all the screen is divided into sub-images; the basis of the preservation process of the correction of various projector display deformation parameters to be displayed on various sub-screen corresponding, respectively, the inverse deformation ; will reverse after the deformation to the corresponding sub-screen display projector. The invention also provides methods for the control device described in this invention. The operation of the invention is simple, computing complexity and low-cost.

The invention discloses a vision-based character string similarity calculation method and similarity judgment method. The character string similarity calculation method comprises the steps of 1) converting each character in two to-be-compared character strings into a corresponding grayscale picture; 2) splicing each row of grayscale values of the picture obtained in the step 1) to obtain a one-dimensional vector of the corresponding picture; 3) calculating the similarity between two one-dimensional vectors corresponding to any two characters in the two to-be-compared character strings, and determining character similarity between the selected two characters according to the similarity; and 4) calculating visual similarity between the two to-be-compared character strings based on the obtained character similarity and character positions. According to the method, difference, presented in a human visual method, of different characters is considered, so that the calculated character string similarity better conforms to a human subjective feeling.

The invention discloses an overall user experience quality assessment method of VR audio and video, particularly relates to an audio and video overall user experience quality assessment method in a virtual reality environment, and belongs to the field of multimedia technology. The implementation method of the invention is as follows: with respect to to-be-measured VR audio and video and referenceVR audio and video, respectively calculating a video quality parameter, an audio quality parameter, an audio and video time parameter and an audio and video space parameter of the to-be-measured VR audio and video through video quality, audio quality and audio and video time interaction and audio and video space interaction models in a first layer; and inputting the four parameters into an overall quality fusion model in a second layer, and finally outputting an overall quality score of the to-be-measured VR audio and video, wherein the space interaction model is implemented by a function relationship designed based on subjective feeling of users. By adoption of the overall user experience quality assessment method disclosed by the invention the influence of the audio and video space interaction on the overall user experience quality of the VR audio and video can be reflected, and the assessment effect better conforms to the subjective feeling of the users.

The invention discloses a full-reference image quality assessment method based on mutual information of Gabor features (MIGF), belonging to the field of image processing. The method specifically comprises the following steps: (I) inputting a reference image and a distorted image; (II) filtering the reference image and the distorted image with two-dimensional Gabor filters respectively to extract Gabor features; (III) building probability distribution functions respectively and computing probability distribution of each pixel and joint probability distribution of the filtered images; (IV) computing the MIGF of the reference image and the distorted image; and (V) building an image quality assessment function MIGF to assess the image quality. The method has the following beneficial effects: the two-dimensional Gabor filters are utilized to extract visual features; the image quality assessment function is built through the mutual information of the reference image and the distorted image; and the assessment results conform to subjective recognition of human visions.

The embodiment of the present invention provides a video coding quality assessment method and a video coding quality assessment device. The method comprises the steps of obtaining the quantification parameters of the strips of a video frame of a video steam and the number of the pixel bytes of the strips of the video frame of the video steam; determining the quantification parameter of a video according to the quantification parameters of the strips of the video frame of the video steam, and determining the content complexity of the video according to the number of the pixel bytes of the strips of the video frame of the video steam; predicting the coding quality of the video according to the content complexity and the quantification parameter of the video. According to the present invention, the content complexity of the video is also considered while the coding quality of the video is predicted, so that the coding quality predicted by a model considering the content complexity of the video accords with the subjective feeling of the human eyes better, and accordingly, the prediction accuracy is improved.

The invention belongs to the field of image processing, and relates to a stereo image quality evaluation method consistent with subjective feeling of human eyes, in particular to a dictionary learning-based stereo image quality evaluation method. An SIFT dictionary and a salient dictionary are trained by utilizing a reference stereo image pair, the SIFT dictionary specifically performs SIFT on thereference stereo image pair, images are represented with SIFT features, and dictionary training is performed by using a feature symbol method and a Lagrangian duality method to obtain the SIFT dictionary; a sparse coefficient matrix is processed to obtain corresponding distorted stereo image quality Q1 and corresponding distorted stereo image quality Q2; and finally the quality Q1 is combined with the quality Q2 to obtain a final quality score Q of a distorted stereo image pair. The method is mainly used for an image processing occasion.

The invention discloses an optimized fusion method based on fuzzy integral, in the HIS space, the intensity component of the spectrum image is decomposed with wavelet multi-layer and acquires the low frequency base band coefficients, acquires the low frequency base band coefficient by decomposing the high resolution image with multi-layer wavelet, integrates the two single factor indexes of spectrum information and space resolution are synthesized with fuzzy integral, carries on pixel fusion with iteration, and carries on character fusion to the high frequency detail of the high frequency sub-band coefficient after being decomposed with wavelet, then carries on correspondent wavelet reverse conversion to the sub-band coefficient and low frequency coefficient, and acquires the new intensity component I', then the fused image may be acquired through HIS reverse conversion. The invention combines the character of HIS fusion method and wavelet fusion method, it realizes high space resolution, and it reduces the deformation of color greatly, and it improves the spectrum information index of fused image.

The invention relates to a tune energy and human ear frequency selectivity based squeal judgment method for speed reducers / changers. The method comprises the steps: firstly, sampling noise signals ofa speed reducer / changer, carrying out A weighting on the noise signals, converting processed signals to log domains, and carrying out Fourier transformation, so as to obtain energy-amplitude frequencycurves; then, calculating feature loudness and critical band rate of the noise signals, calculating frequency band acuity of each feature loudness, and subjecting frequency band acuity to normalizingtreatment; subjecting the energy-amplitude frequency curves to frequency interval Gauss smoothing processing to obtain background noises, and selecting candidate tunes; calculating a frequency band acuity weighting factor according to a corresponding relationship between the critical band rate and frequency, and carrying out weighting, so as to obtain the energy-amplitude frequency curves of thenoise signals; calculating tune energy in bandwidth of the tunes, background noise energy and difference between the tune energy and the background noise energy, and carrying out primary and secondarysequencing on abnormal noise frequencies affecting squeal. According to the method, a low-frequency region becomes a non-sensitive region, and a high-frequency region highlights acuity attribute, andthus, a subjective sensation of human ears to squeal is effectively simulated.

The invention discloses a low-bit-rate video quality detection method based on blocking effects and noises. The method comprises the following steps of: 1) obtaining a reference video and a distortion video from a public video library, and respectively converting the reference video and the distortion video into corresponding images by frames, namely a reference video frame and a distortion video frame; 2) carrying out blocking effect detection to the images; 3) carrying out noise detection to the images; and 4) fusing a blocking effect operator and a noise operator to improve an SSIM (structural similarity) algorithm. The method has the beneficial effect that the method fuses blocking effect concealment, noise concealment and the like to more conform to the human vision feeling. The simpleness and the high efficiency of the SSIM algorithm are ensured, meanwhile, the method is improved based on the human eye characteristic to more conform to the subjective feeling of the human eye.

The invention relates to the technical field of image quality evaluation, particularly provides an image quality evaluation method and device and a computer readable storage medium, and aims to solve the technical problem of how to improve the accuracy of image quality evaluation. For the purpose, the method provided by the embodiment of the invention comprises: obtaining the pixel feature of each pixel in a to-be-evaluated image, and obtaining a first image quality evaluation index value of the to-be-evaluated image according to the pixel feature; dividing the to-be-evaluated image into a plurality of image blocks, and performing image quality analysis on each image block by adopting a preset image quality evaluation model to obtain a second image quality evaluation index value of the to-be-evaluated image; obtaining a final image quality evaluation index value according to the first image quality evaluation index value and the second image quality evaluation index value; and performing image quality evaluation on the to-be-evaluated image according to the final image quality evaluation index value. Through the above steps, the accuracy of image quality evaluation can be improved.

The invention belongs to the field of image processing, and discloses an underwater image data set construction method and system, and the method comprises the steps: constructing underwater fuzzy image data sets of various types, and building a background light data set based on a manual marking method; meanwhile, evaluzting the underwater image quality based on four-dimensional area evaluation indexes of image chromaticity, contrast ratio, gradient and sharpness, so that constructing an underwater high-definition image data set. Based on human subjective discrimination, the accuracy of output underwater background light can be ensured; the constructed underwater blurred image data set and the background light data set thereof can be used as a data source and background light estimation reference in underwater image restoration research; the underwater high-definition image data set construction method is simple and rapid, the constructed underwater image data set can provide trainingsamples for an underwater image quality enhancement algorithm based on deep learning, and the underwater image quality evaluation method can be used for underwater image quality evaluation.

The invention discloses an influence evaluation system construction method and system and influence evaluation method. The influence evaluation system construction method comprises the following stepsof: obtaining a set quantity of history data of at least one influence index in at least one influence dimensionality; respectively converting the obtained data of each influence dimensionality intodecibel numbers; and constructing an influence index model through a dimensionality reduction algorithm according to the decibel numbers. According to the influence evaluation system construction method, comprehensive influences of media objects such as articles, events and media are scientifically and quantitatively assessed through basic data such as reading numbers, like numbers, forwarding numbers and comment numbers; besides ensuring mathematically optimum, the method does not depend on preferences of calculation formula designers; all the weights are calculated according to large-scale data, so that objectivity is provided; and the finally calculated network media influences, namely the influence indexes, are based on specific media with high influences, so that values of the influence indexes have clear physical meanings.

The present invention provides a presswork ink spot detection method and apparatus. The method includes the following steps that: the original images of presswork and a reflectivity calibration sample, which are acquired by a scanner, are acquired; the original images of the presswork are sampled, so that a plurality of ink spot images to be detected are obtained; wavelet decomposition, high-frequency de-noising processing and image reconstruction are performed on the plurality of ink spot images, so that smoothed ink spot images can be obtained, and the discrete matrix gray value of the smoothed ink spot images of the presswork is determined; the original images of the presswork are compared with the original images of the reflectivity calibration sample, so that the average reflectivity of the presswork can be determined; and the discrete matrix gray scale value variation coefficient of the smoothed ink spot images of the presswork is determined according to the average reflectivity and the discrete matrix gray value. According to the presswork ink spot detection method and apparatus of the invention, human eye-based sensitivity comparison is used in combination, a digital image processing technology is utilized, and therefore, a detection result can better accord with the subjective perception of human eyes, the relevance of the detection result and human eye visual evaluation can be improved, and simple, high-precision and low-cost presswork ink spot detection can be realized.

The invention provides a quality estimation method of exit-entry digital portrait photos. The quality estimation method comprises the following steps: the format size and file size detection of each digital portrait photo, face detection, eye positioning, face horizontal centering detection, the exposure detection of global luminance deviation and partial luminance deviation, the clearness detection of contrast evaluation and ambiguity evaluation and the color detection and the background detection for detecting the skin naturalness in an Lab color space and judging whether colors in an imageface area are qualified or not. The quality estimation method has the advantages that a quality estimation module of the digital portrait photos, specific quality detection items, detection flows, estimation indexes and estimation methods are established, which is different from the prior quality estimation method of the digital portrait photos relying on subjective estimation. The invention judges the quality of each digital portrait photo without relying on a standard image, obtains the judging result in accordance with the subjective feeling of people and the requirements of a face recognition system, thereby favorably satisfying the business requirements for managing the exit-entry digital portrait photos.

The invention discloses a structural tensor total variation-based image and video super-resolution method. The method includes the following steps that: an image degradation model is established, wherein the image degradation model includes a blurring part and a downsampling part and is expressed as an expression described in the descriptions of the invention, wherein x represents an original high-resolution image, B represents a blurring kernel, a symbol, described in the descriptions, represents a convolution process, D represents downsampling of a corresponding multiple, and y represents a generated low-resolution image; a super-resolution model is established, wherein the super-resolution model includes an image interpolation part and a de-blurring part and is expressed as an expression described in the descriptions of the invention, wherein x<^> represents a restored high-resolution image, B represents a blurring kernel, a symbol, described in the descriptions, represents a deconvolution process, D<-1> represents an interpolating process of a corresponding multiple, and y represents a low-resolution image; and a structural tensor total variation regularization constraint-based super-resolution model is constructed and is expressed as an expression described in the descriptions of the invention, wherein Mu is a parameter that can be adjusted, so that the strength of the constraint of a regularization term can be realized. With the structural tensor total variation-based image and video super-resolution method adopted, over-smoothing and an edge step effect caused by a condition that a total variation model directly constrains an image, are avoided, and a processing effect obtained by using the method of the invention more accords with the subjective feeling of eyes.

The invention provides a structure-enhanced semi-supervised online map generation method, and relates to the technical field of online maps. The structure-enhanced semi-supervised online map generation method comprises the following specific steps: acquiring two image resources of a remote sensing image and an instant map by using an existing remote sensing image and a map data set or through an API provided by a Google map, and carrying out standardization processing, so that a mapping relation of partial samples exists between the two acquired image resources; and according to whether pairing is carried out or not, dividing the samples into pairing samples and non-pairing samples through a file organization structure, and constructing a semi-supervised training data set. According to the method, a semi-supervised training method following circulation consistency is established, available resources are more fully applied, structural characteristics of CyclGAN and Pix2Pix are combined, a semi-supervised learning strategy of S2OM is created, paired samples and non-paired samples are learned in stages, model learning is more sufficient, and the map generation error rate is lower.

The invention provides a combined optimizing and training method for the multi-dictionary learning of image sparse representation, which belongs to the multi-media communication and image data processing field, and is characterized by: regarding the non-zero element in the singular value matrix after the singular value decomposition of the gradient matrix of the training image unit as the energy value of the gradient direction; according to set the energy value parameter threshold, dividing the image unit into various isotropic images and various anisotropic images; learning successively a shared dictionary and a specialized dictionary; optimizing the factors to minimize the objective function such as the residuals of the isotropic and anisotropic images after the sparse representation, the autocorrelation and the degree of cross-correlation of each dictionary, and the regularization of nonzero element; in the optimization process, using the orthogonal matching and tracking algorithm to optimize A0 and Ak, and then optimizing D0 and Dk by the gradient descent algorithm; and in keeping the parameters to be optimized, regarding the items not involving the parameters as constants. When the method is used for image compression, the details could be well preserved, a low distortion rate is achieved and a good image quality could also be achieved.

The invention discloses a visual-based character string similarity calculation method and a similarity judgment method. The method for calculating the character string similarity of the present invention is: 1) converting each character in the two character strings to be compared into a corresponding gray scale image; 2) splicing the gray value of each line of the image obtained in step 1) , to obtain the one-dimensional vector of the corresponding picture; 3) for any two different characters in the two character strings to be compared, calculate the similarity of the two one-dimensional vectors corresponding to these two characters, and determine the character similarity of the selected two characters according to the similarity 4) Calculate the visual similarity of two character strings to be compared based on the character similarity and character position obtained above. This method takes into account the differences that different characters present in human visual methods, and the calculated string similarity is more in line with human subjective feelings.

The invention discloses a full-reference image quality assessment method based on mutual information of Gabor features (MIGF), belonging to the field of image processing. The method specifically comprises the following steps: (I) inputting a reference image and a distorted image; (II) filtering the reference image and the distorted image with two-dimensional Gabor filters respectively to extract Gabor features; (III) building probability distribution functions respectively and computing probability distribution of each pixel and joint probability distribution of the filtered images; (IV) computing the MIGF of the reference image and the distorted image; and (V) building an image quality assessment function MIGF to assess the image quality. The method has the following beneficial effects: the two-dimensional Gabor filters are utilized to extract visual features; the image quality assessment function is built through the mutual information of the reference image and the distorted image; and the assessment results conform to subjective recognition of human visions.