168results about How to "Increase salience" patented technology

The method and preprocessor enhances the intelligibility of narrowband speech without essentially lengthening the overall time duration of the signal. Both spectral enhancements and variable-rate time-scaling procedures are implemented to improve the salience of initial consonants, particularly the perceptually important formant transitions. Emphasis is transferred from the dominating vowel to the preceding consonant through adaptation of the phoneme timing structure. In a further embodiment, the technique is applied as a preprocessor to a speech coder.

The invention discloses a speech emotion recognition method. The method includes the steps that firstly, a speech signal is converted into a spectrogram to serve as initial input; secondly, a deep convolutional neural network is trained to automatically extract emotion features; thirdly, a stack type auto-encoder is trained for each kind of emotions, and all the stack type auto-encoders are fused to automatically construct membership functions of an emotion fuzzy set; fourthly, the features obtained in the second step are subjected to feature optimization by means of the fuzzy optimal theory in the third step; fifthly, emotion classification recognition is conducted by means of a Softmax classifier. The method takes abstract fuzzy properties of speech emotion information into consideration, the extracted emotion features are subjected to selective fuzzy optimization to improve the significance of the features, fuzzy membership functions in the fuzzy theory are automatically constructed by means of the concept of deep neural network layer-by-layer training, and the problem that the proper membership functions in the fuzzy theory are difficult to select and determine is solved.

The invention discloses a deep residual network-based semantic mammary gland molybdenum target image lump segmentation method. The method comprises the following steps of: labelling pixel categories of lumps and normal tissues corresponding to a collected mammary gland molybdenum target image so as to generate label images, and dividing the mammary gland molybdenum target image and the corresponding label images into training samples and test samples; preprocessing the training samples to form a training data set; constructing a deep residual network, and training the network by utilizing thetraining data set, so as to obtain a deep residual network training model; after a to-be-segmented mammary gland molybdenum target image lump is preprocessed, carrying out binary classification and post-processing on a pixel of the to-be-segmented mammary gland molybdenum target image by utilizing the deep residual network training model, and outputting lump segmentation image to realize semanticsegmentation of the mammary gland molybdenum target image lump. The method is capable of effectively improving the automatic and intelligent levels of mammary gland molybdenum target image lump segmentation, and can be applied to the technical field of assisting radiologists to carry out medical diagnosis.

The invention discloses a brain image segmentation method and system based on a saliency learning convolution nerve network; the brain image segmentation method comprises the following steps: firstly proposing a saliency learning method to obtain a MR image saliency map; carrying out saliency enhanced transformation according to the saliency map, thus obtaining a saliency enhanced image; splitting the saliency enhanced image into a plurality of image blocks, and training a convolution nerve network so as to serve as the final segmentation model. The saliency learning model can form the saliency map, and said class information is obtained according to a target space position, and has no relation with image gray scale information; the saliency information can obviously enhance the target saliency, thus improving the target class and background class discrimination, and providing certain robustness for gray scale inhomogeneity. The convolution nerve network trained by the saliency enhanced images can be employed to learn the saliency enhanced image discrimination information, thus more effectively solving the gray scale inhomogeneity problems in the brain MR image.

The present invention varies the highlight color of a highlighted display area such that the salience of the highlighted display area remains constant to a user as the highlighted display area varies. A perceptual highlight color is determined based upon the total amount of highlighted display area and the background color of the non-highlighted display area and applied to text selected to be highlighted by a user as well as to text already highlighted on the display. In other embodiments, the perceptual highlight color can be based upon the total amount of highlighted display area and the text color of the text in the highlighted display area, and the total amount of the highlighted display area, the background color of the non-highlighted display area, and the text color.

A method on a computing device is provided for enhancing the memory and cognitive ability of an older adult by requiring the adult to differentiate between rapidly presented stimuli. The method utilizes a sequence of phonemes from a confusable pair which are systematically manipulated to make discrimination between the phonemes less difficult or more difficult based on the success of the adult. The manipulation includes processing of the consonant and vowel portions of the phonemes by emphasizing the portions, and / or by stretching the portions. Further processing includes separating the consonant and vowel portions by time intervals. As the adult improves in their auditory processing, the discriminations are made progressively more difficult by reducing the amount of processing to that of normal speech.

The invention belongs to the technical field of biology and discloses a group of circRNA markers for early diagnosis of lung adenocarcinoma and application thereof. The circRNA marker disclosed by the invention comprises four kinds of circRNA, namely circRNA-100323, circRNA-69774, circRNA-403738 and circRNA-404449. The invention has the beneficial effects that a group of novel molecular bio-markers with diagnostic values on lung adenocarcinoma are provided. According to verifications of clinical sample cases, the group of circRNA markers have high diagnostic specificity and sensitivity on early diagnosis of lung adenocarcinoma.

The invention discloses a video compression method based on a convolutional neural network and HEVC compressed domain significant information. The video compression method based on the convolutional neural network and the HEVC compressed domain significant information improves and strengthens the HEVC from two aspects of a video saliency algorithm based on an attention mechanism and a perception preferred video compression algorithm; in the aspect of the video saliency, the method combines motion estimation results of each CU in the HEVC compression process on the basis of the convolutional neural network to adaptively integrate the two in a dynamic condition, so as to complete the saliency detection of an input video; in the aspect of the perception preferred video compression algorithm,a corresponding QP is selected according to a significant value of the CU, so that the CU with a higher significant value can be encoded with a smaller QP, and the saliency feature of the current CU block is incorporated into a traditional rate distortion calculation method, so as to achieve the purpose of perception priority. The video compression method based on the convolutional neural networkand the HEVC compressed domain significant information reduces the perception redundancy of the video and thus obtains a better compression effect.

The invention relates to the technical field of biology, and in particular relates to a group of LncRNA biomarkers for diagnosing human lung adenocarcinoma and human colorectal cancer, related kits and application thereof. The group of LncRNA biomarkers for diagnosing human lung adenocarcinoma and / or human colorectal cancer comprises uc001gzl.3, ENST00000434223, uc004bbl.1, ENST00000540136 and NR_034174. The LncRNA biomarkers for diagnosing the human lung adenocarcinoma and / or human colorectal cancer can be used for distinguishing human early lung adenocarcinoma and human colorectal cancer, the AUC of the biomarkers for distinguishing early lung adenocarcinoma from normally paired lung tissues can reach 0.978, and the sensitivity and the specificity are respectively 92% and 98%. The AUC of the biomarkers for diagnosing a colorectal cancer sample can reach 0.963, and the sensitivity and the specificity are respectively 94.4% and 88.9%.

The invention discloses a current signal-based motor bearing fault diagnosis method. According to the method, motor current history signals under different bearing fault states are obtained in a modeltraining stage; harmonic signals are extracted from the current history signals and fundamental frequencies and harmonic waves are eliminated from original current history signals to obtain residualsignals; time domain analysis and frequency domain analysis are carried out on the residual signals to extract bearing fault feature indexes; training is carried out on the basis of the bearing faultfeature indexes and bearing fault types to obtain a bearing fault diagnosis model; and carrying out processing same as the model training stage on a to-be-diagnosed motor current real-time signal to obtain a bearing fault feature index, and the bearing fault feature index is input into the trained bearing fault diagnosis model to carry out mode recognition so as to diagnose a bearing fault state.The method is capable of solving the technical problem that mistaken and false information is caused as existing motor bearing fault diagnosis methods are difficult to extract fault feature signals from current signals with low signal to noise ratios.

A method on a computing device is provided for enhancing the memory and cognitive ability of an older adult by requiring the adult to listen to two or more aurally processed syllables, presented serially, view the syllables graphically, and then designate the order in which the syllables were aurally presented. A number of trials are presented to the adult. As the adult correctly determines the syllable order in trials, the difficulty of the trials is increased by increasing the number of syllables presented, and by reducing the amount of processing that is applied to the syllables.

A method on a computing device for enhancing the memory and cognitive ability of an older adult by requiring the adult to differentiate between rapidly presented stimuli. The method utilizes a sequence of phonemes from a confusable pair which are systematically manipulated to make discrimination between the phonemes less difficult or more difficult based on the success of the adult, such as processing the consonant and vowel portions of the phonemes by emphasizing the portions, stretching the portions, and / or separating the consonant and vowel portions by time intervals. As the adult improves in auditory processing, the discriminations are made progressively more difficult by reducing the amount of processing to that of normal speech. Introductory phonemes may each include a blend of a formant-synthesized phoneme and an acoustically naturalistic phoneme that substantially replicates the spectro-temporal aspects of a naturally produced phoneme, with the blends progressing from substantially natural-sounding to substantially formant-synthesized.

A method on a computing device is provided for enhancing the memory and cognitive ability of an older adult by requiring the adult to listen to two or more aurally processed syllables, presented serially, view the syllables graphically, and then designate the order in which the syllables were aurally presented. A number of trials are presented to the adult. As the adult correctly determines the syllable order in trials, the difficulty of the trials is increased by increasing the number of syllables presented, and by reducing the amount of processing that is applied to the syllables.

The invention relates to the technical field of cosmetics, and in particular relates to a composition and an application of the composition in the preparation of the moisturizing cosmetic. In a formula provided by the invention, the components cooperate with one another, and jointly achieve the effect of improving moisture content of skin and reducing moisture dissipation of skin. The effect is significantly improved when being compared with the effects achieved by independently using the certain component or the several components, which illustrates that the components compounded in a proper proportion generate the synergistic effect, and thus the moisturizing effect is improved.

The invention discloses a plane intensive parking garage operating method and equipment, and relates to an operating method and equipment for a mechanical translation parking system, by which vehicles can be moved under the condition that a same closed rectangular plane is full of arranged vehicles. In the same rectangular plane, a plurality of vehicle carrying plates (1) are connected end to end and are arranged in the manner of linear rows and ranks, the number of either of the rows and the ranks is 2, the number of the other is 2 or an integer larger than 2, and only one vacancy (2) is reserved in the range of the same rectangular plane. In the method provided by the invention, the vehicle carrying plates are moved by a driving mechanism to slide in the rectangular plane, and are moved one by one from the vehicle carrying plate close to the vacancy. In a closed alley where only two vehicles are allowed to be arranged, the method and equipment provided by the invention can realize the remarkable effect of increasing parking space; and in a smaller basement, the method and equipment also can realize the remarkable effect of increasing parking space.

The invention discloses a sea-surface target saliency detection method based on a convolutional neural network. The sea-surface target saliency detection method comprises the steps of: s1, inputting an image I to be subjected to saliency detection into a trained convolutional neural network structure to obtain a multi-scale feature map of the image, and performing saliency calculation on the multi-scale feature map to obtain a corresponding first salient map; S2, meanwhile, performing color space conversion on the image I to be subjected to saliency detection to obtain a channel image, and performing saliency calculation in a spatial domain on the channel image to obtain a corresponding second salient map; S3, and fusing the first salient map and the second salient map to obtain a final salient map. The sea-surface target saliency detection method can improve the problems of insufficient feature extraction and poor robust detection result existing in the present saliency detection, sothat the saliency of a saliency region is significantly improved, and the precision of the salient result is improved.

A method and instrument for detecting and analyzing the predication degree index of cardiacrhythm variability features that the predication degree is used as the index of detecting and analyzing cardiac rhythm variability. Said method includes such steps as using dynamic electrocardiograph to record cardioelectric signals, automatically recognizing dynamic cardioelectric RR gap, forming time sequence RRi, and calculating non-linear predication degree PI.

The invention relates to a weak and small object recognition method based on a deep neural network, and the method is an application of the deep neural network at the aspect of computer vision. Aimingat the problems that an existing neural network is not high in recognition precision of a weak and small object and is slow in recognition speed, the invention proposes a novel deep neural network structure, and the structure focuses on the extraction of the features of the weak and small object and the accurate description of the small-scale image features through the increase of the image datadepth and the repeated extraction of features at the same scale, thereby achieving the accurate recognition of the weak and small object.

The invention discloses an image saliency detection method and device, a storage medium and a processor. The image saliency detection method comprises the steps of performing super-pixel segmentation on a depth map of a target image to obtain a plurality of super-pixels of the target image; performing clustering processing on each super-pixel to obtain a first number of categories of each super-pixel, wherein the clustering center of a category containing the largest number of pixels is the target clustering center of each super-pixel in the first number of categories; and carrying out a first preset algorithm on the target clustering center of each super-pixel to obtain a depth saliency map of the target image. The accuracy of saliency detection for an image is improved according to the invention.

The invention discloses a (+)-(3R,4R)-[[2S-[[4-(3-hydroxyphenyl-3,4- dimethyl-1-nipecotic)-methyl]-1-oxo-3-phenylpropyl]amido]benzyl cetate compound and its process for preparing and a new technology for preparation of Mope by the said compound. The new technique of this invention changes the íŒone chainíŒ synthesis of the original patent method to the separate í‹two chainsíŒ preparation, by which the total compound steps are dramatically shortened, the compound cost is reduced to about one-half of the original patent method, and the method is more competitive on the market in price. Furthermore,the reaction condition is moderate, the operating procedure is simple, each reaction step is no need of given condition such as low temperature and absolute water-free, so it is more suitable for industrial production.

Embodiments of the invention disclose an image processing method and apparatus, and a server. The method comprises the following steps of obtaining a to-be-processed human face image; inputting the human face image to a convolutional neural network model with a loss function, wherein the loss function directionally screens and increases a between-class distance after image classification accordingto a preset expectation; and obtaining classification data output by the convolutional neural network model, and according to the classification data, performing content understanding on the human face image. The new loss function is established on the convolutional neural network model and has the effect of screening and increasing the between-class distance after the image classification; and the between-class distance of the classification data output by the convolutional neural network model obtained by training through the loss function is increased, so that the between-class distance inan image identification process is increased, the saliency of difference between images is remarkably improved, the image comparison accuracy is remarkably improved, and the security of applying theimage processing method is effectively guaranteed.

The invention provides a three-dimensional point cloud data instance segmentation method and system in an automatic driving scene. The method comprises steps of carrying out the preliminary recognition and division of an outdoor street scene through the spatial position information of a target object, and forming a point cloud visual column of an interested region; visual column point clouds containing objects and negative sample visual column background point clouds distributed in the same way are extracted from the point cloud visual columns of the region of interest to form a visual columnpoint cloud data set; and extracting high-dimensional semantic feature information of an object contained in each visual column point cloud in the visual column point cloud data set, and meanwhile, introducing a multi-classification focus loss function with a weight to obtain a category to which each point cloud in the visual column belongs, thereby realizing instance segmentation of the point cloud data. According to the three-dimensional point cloud data instance segmentation method in the automatic driving scene, target detail feature expression can be effectively enhanced so that the prediction capability of point cloud difficult samples can be enhanced and the performance of point cloud instance segmentation in the automatic driving scene can be enhanced.

A method on a computing device for enhancing the memory and cognitive ability of an older adult by requiring the adult to differentiate between rapidly presented stimuli. The method utilizes a sequence of phonemes from a confusable pair which are systematically manipulated to make discrimination between the phonemes less difficult or more difficult based on the success of the adult, such as processing the consonant and vowel portions of the phonemes by emphasizing the portions, stretching the portions, and / or separating the consonant and vowel portions by time intervals. As the adult improves in auditory processing, the discriminations are made progressively more difficult by reducing the amount of processing to that of normal speech. Introductory phonemes may each include a blend of a formant-synthesized phoneme and an acoustically naturalistic phoneme that substantially replicates the spectro-temporal aspects of a naturally produced phoneme, with the blends progressing from substantially natural-sounding to substantially formant-synthesized.

The invention discloses an SAR image registration method based on multi-scale image block characteristics and sparse expression, and mainly solves the problem that an existing registration method is poor in effect when applied to SAR image registration. The SAR image registration method comprises the steps of: 1) inputting two SAR images, selecting any one as a reference image, and using the other one as an image to be registered; 2) selecting characteristic points of the reference image; 3) utilizing the multi-scale image block characteristics to construct characteristic point descriptors of the reference image and the image to be registered; 4) establishing matching point pairs between the reference image and the image to be registered; 5) removing abnormal points in the matching point pairs; 6) establishing an affine transformation model according to the finally obtained matching point pairs, adopting a least square method to obtain geometric deformation parameters, and obtaining a registration result. Compared with the prior art, the robustness to spot noise is improved, the accuracy and the matching precision of the matching point pairs are improved, and the SAR image registration method can be applied to image fusion and change detection.

The invention discloses a fine-grained image classification method based on a generative adversarial network and an attention network, and the method comprises the steps: determining an image classification category, and building a training image set of a corresponding category; designing a deep attention convolution network for image fine-grained classification, wherein the network comprises fourparts, namely a VGG16 full convolution layer, SS attention region generation, a spatial pyramid pooling layer ROI pooling layer and an overall and local feature combined classification full connection layer; designing a structure of a generative adversarial network DAC-GAN, a generative network and a discrimination network; training a DAC-GAN network by using the training sample set, and storinga discriminant network model; and carrying out classification prediction on the image types by using the discrimination network model. According to the invention, the accuracy of the image classification network is improved, and the problem of insufficient data of a small sample size is solved.

The invention discloses a polycrystalline packaging unit, comprising a substrate (1), at least one hollow fixed frame (4), a plurality of LED crystal grains (6) and a packaging colloid (5), wherein the substrate (1) has high heat radiation, is made of metals to be in a shape of thin plate, and is provided with an insulating layer (10) on the surface, a circuit layer (2) arranged on the top of the insulating layer (10) to be connected with a power cord, and a protective layer (3) which is arranged on the top of the circuit layer (2) to be used for covering circuits and has insulating performance; the fixed frame (4) is arranged on the surface of the protective layer (3) and is smaller than the substrate (1); the LED crystal grains (6) are arranged in the fixed frame (4) and is electrically connected with the circuit layer (2); and the packaging colloid (5) is arranged in the fixed frame (4), covers the LED crystals (6) and has light transmission. The polycrystalline packaging unit has the advantages of simple manufacture, favorable heat radiation and high brightness.

The invention discloses a blank hub drilling center accurate positioning method based on machine vision. The advantage of the method is characterized by, through superpixel segmentation technology, carrying out clustering segmentation on visual images; calculating gradient energy of each cluster superpixel for the segmented images to enable significance of hub edge to be enhanced greatly and allow a hub image to be segmented out accurately; then, according to the segmented hub image, obtaining an initial circle center and radius, and with the initial circle center being as a reference circle center, carrying out scanning to obtain edge scanning data; removing noise interference points of the obtained scanning data; and carrying out least square circle fitting to realize accurate positioning required by blank hub drilling positioning.

The invention provides a target saliency detection method for rainy day images. The method comprises the following steps: S1, extracting visual saliency based on luminance and color features, calculating luminance saliency map Slc; S2, extracting visual saliency based on color difference features, and calculating color difference saliency map Scv; S3, extracting the visual saliency based on the dark channel, and calculating the saliency map Sd of the dark channel; S4, mixing the brightness saliency map Slc, the color difference saliency map Scv and the dark channel saliency map Sd to obtain the final saliency map Sfinal, Sfinal=Slc. * Scv-Sd. In the invention, the brightness and color characteristics, the color difference characteristics and the dark channel characteristics of the target in the rainy day image are utilized to extract the saliency characteristics of the target in the rainy day image and construct a target saliency detection model, thereby providing an effective preprocessing method for the image sharpening and the detection of the target in the rainy day.

The invention discloses a saliency detecting method applied to static image human segmentation. The method comprises the following steps of performing superpixel segmentation on a static image to be detected; performing face detection on the image after the superpixel segmentation to acquire a face area; performing skin color detection on the face area to acquire skin color information; performing color uniqueness calculation and color spatial distribution calculation according to the skin color information to acquire a color uniqueness value with the skin color information fused therein and a color spatial distribution value with the skin color information fused therein; performing saliency calculation according to the acquired color uniqueness value and the color spatial distribution value to acquire a saliency map used for static image human segmentation. Based on a conventional saliency detecting method, the method provided in the invention makes the skin color well detected by making the skin color information fused in the color uniqueness calculation and color spatial distribution calculation. The human skin color saliency is enhanced. Better and more accurate static image human segmentation effects are achieved. The method can be widely applied to the image processing field.

The invention provides an automatic detection method of a salient object based on salience density and edge response, and relates to a method for automatically detecting the salient object, solving the problems of the convectional salient object detection method that only one attribute that is the salience is utilized, but the edge attribute of the salient object is not taken into account, therefore, the detection accuracy of the salient object is relatively low. The automatic detection method of the salient object based on the salience density and the edge response comprises the following steps of: calculating and generating a salient map S of an input map according to the regional salience calculation method in combination of the global color comparison and the color space distribution; generating an edge response map E on the salient map S by utilizing a group of Gabor filters; efficiently searching a global optimal sub-window containing the salient object in the input map by utilizing the maximized branch-and-bound algorithm of the salience density and the edge response; adopting the obtained optimal sub-window as the input; initializing the GrabCut graphic cutting method; carrying out the GrabCut graphic cutting method; and automatically extracting the salient object with a good edge. The automatic detection method is applicable to the image processing field.