82 results about "Classifier fusion" patented technology

The present invention includes a system and method for automatically extracting the demographic information from images. The system detects the face in an image, locates different components, extracts component features, and then classifies the components to identify the age, gender, or ethnicity of the person(s) in the image. Using components for demographic classification gives better results as compared to currently known techniques. Moreover, the described system and technique can be used to extract demographic information in more robust manner than currently known methods, in environments where high degree of variability in size, shape, color, texture, pose, and occlusion exists. This invention also performs classifier fusion using Data Level fusion and Multi-level classification for fusing results of various component demographic classifiers. Besides use as an automated data collection system wherein given the necessary facial information as the data, the demographic category of the person is determined automatically, the system could also be used for targeting of the advertisements, surveillance, human computer interaction, security enhancements, immersive computer games and improving user interfaces based on demographic information.

The invention discloses a system and a method for filtering sensitive webpage, which is based on multi-classifier fusion. The processing object is a webpage, and the processing result is whether the webpage contains sensitive content, which may be pornography, reaction, violence and other unhealthy Internet contents harmful to society. The system comprises a data stream obtaining and preprocessing unit, an image and text stream filtering unit and an information fusion unit of image filter and text filter, by the cooperation of multiple classifiers, the system acquires source code of a webpage by using the URL of the webpage, a text and an image are separated at preprocessing stage to obtain text information and effective image information; an input webpage is divided into three modes by decision tree algorithm; the webpage is recognized by using a consecutive text classifier, a discrete sensitive text classifier and an image classifier, the output result recognized by the classifiers is fused and calculated, then a judge factor is given, and the final result is returned to a browser.

The invention discloses a garment identification method and system for a low-resolution video. The garment identification method is based on a space-time classifier fusion technology and specifically comprises the following steps of: extracting foreground images from a video stream, and extracting contour information on moving objects in the foreground images; carrying out identification on moving human objects according to the extracted contour information; carrying out multi-point feature identification treatment on different subblocks of the same human object in an image of the same frame in the video stream, and carrying out voting judgment on an identified result; and carrying out voting judgment on identified results for the same human object in multi-frame images in the video stream, and finally, determining the categories of the garments for the moving human objects. With the adoption of the space-time classifier fusion technology-based method disclosed by the invention, the categories of the garments and the identities of the human objects are finally determined through carrying out judgment on the garment features of the same moving object in a plurality of video frames based on motion detection, human body identification and garment identification, and thus, the aim of high-efficiency, high-quality and high-accuracy identification of the identities and the garments is achieved.

The invention relates to a brain emotion identification method based on a multi-classifier fusion model constructed via a hierarchical mechanism. Multi-channel emotion brain data is collected and analyzed by brain preprocessing, feature extraction and weight measuring channel selection to construct an emotion brain feature matrix. Channel division is carried out on the emotion brain feature matrix according to electrode positions, optimized feature selection integration is carried out on each channel, and single-emotion classification models are formed. The difference and accuracy of the classification models in solving the same emotion identification problem serve as evaluation criteria, an optimal single-emotion classification model is selected from each channel, and a classifier set to be fused is obtained. An emotion identification fusion model is constructed on the basis of a weighted voting method by utilizing classification errors of the optimal single-emotion classification models serve as weights. According to the invention, multi-classifier fusion is used to solve problem that a high emotion identification rate is hard to obtain in a brain sample space.

The invention relates to a face identification method and a face identification system based on fusion of multiple classifiers. The method comprises following steps that a second classifer acquires a left-handed face image with a pose range of [-90, -15], a third classifier acquires a front face image with a pose range of [-15, +15], a fourth classifier acquires a right-handed face image with a pose range of [+15, +90], and comparing faces in the face images; fusing the face image with the comparison results of the same compared person in a pose module database to acquire an identification result, the identified relevant information of the compared person is displayed or the left-handed face image, the front face image and the right-handed face image are stored into the pose module database. Due to the adoption of the method and the system, a first comparison result, a second comparison result and a third comparison result can be effectively fused, the variation situation of the face poses in an application environment can be effectively processed in real time, and the accuracy and the robustness for identifying the multi-pose face can be improved.

The present invention is method of recognizing and shielding the sex part in fetus image includes the following steps: 1. processing the original fetus image, finding out the edge contour of the sex part, and determining at least one initial locating point of suspected sex part; 2. recognizing the initial locating point by means of the fetus sex image database and recognizing system with multiple classifier fusion structure to judge the fetus sex part; and 3. shielding the fetus sex part with shielding image. The present invention makes it possible to avoid illegal fetus sex identification in ultrasonic apparatus effectively.

The invention provides an image classification method and apparatus based on multi-core learning classifier fusion. The method comprises steps of: S1, establishing a sample library which includes different types of samples; S2, performing feature extraction on the different types of samples and obtaining kernel functions corresponding to the different types of samples according to feature extraction results; S3, synthesizing the obtained kernel functions to establish a multi-core model; S4, training the multi-core model to obtain a plurality of classifiers; S5, using an Adaboost algorithm to assign different weights to the plurality of classifiers obtained in the S4 so as to fuse the plurality of classifiers to obtain a target classifier; and S6 classifying the images to be classified by using the target classifier to obtain a classification result. The image classification method of the present invention can improve the accuracy of image classification.

The invention discloses a recommendation method based on singular value decomposition and classifier combination. The recommendation method comprises the steps of computing average score and probability distribution of an item through data preprocessing; training a singular value decomposition model through a stochastic gradient descent method, computing an entropy set of a scored item set of the user in item classification through a computing method of entropy, and determining an uncertainty critical value of the item; and comparing and predicting uncertainty and critical value of the item to determine whether to use a classifier, and recommending N items with highest scores in all non-scored items of the user through a Top-N method. According to the method, individual recommendation is produced on the basis of analysis of historical score data of the user; predicting score of a designated item i is acquired through a singular value decomposition algorithm, information entropy of the item for each user is calculated so as to determine whether to classify, and final prediction score of the item is acquired through the classifier, so that the accuracy of the recommendation method is improved.

The invention belongs to the technical field or image processing and pattern recognition, and discloses a gender identification method based on a facial image. According to the gender identification method, a classifier fusion method that a classifier used for training features of local organs (for example, the five sense organs) on a human face respectively and a classifier of overall features of the human face are combined is used for establishing a fusion classifier for carrying out gender identification, so that the identification accuracy rate is improved; and since a two-dimensional principal component analysis (2DPC) method is used for carrying out dimensionality reduction on the facial image, a two-dimensional linear discriminant analysis (2DLDA) is used as a classifying method, on the basis that the detection accuracy is guaranteed, the calculation amounted is reduced, and the training speed and the detecting speed are increased.

The invention discloses a video target detection method based on machine learning. The method comprises the steps that (1) for an input video, an SSD target detection algorithm is adopted to obtain ato-be-tracked target detection box, and a bounding-box is marked on an image to determine a tracking target; (2) two tracking methods are adopted for each frame of the input video, wherein one tracking method is a light stream tracking algorithm, a tracking point of the next frame is predicted according to a probability, and the tracking point of the next frame is precisely determined through a Euclidean distance and a set threshold value; and the other tracking method is to adopt a full-convolutional neural network and extract high-layer features and low-layer features in the neural network for separate convolution, finally the features are fused into a feature graph through a classifier, and therefore the tracking point of the next frame is precisely determined; and (3) HOG features of the light stream tracking result and the full-convolutional neural network tracking result are extracted, validity discrimination is performed on the two results through a support vector machine (SVM),and the target position of the next frame is determined finally.

The invention provides a fuzzy integration multiple classifier integration-based uterine neck cell image identification method and device. The uterine neck cell image feature identification method comprises the following steps: in step S10, a single uterine neck cell image is subjected to preprocessing operation, edges of cytoplasm and cell nucleuses can be highlighted on the cell image, and a uterine neck cell image background can be removed; in step S20, the preprocessed uterine neck cell image is subjected to improved CV model segmenting operation; in step S30, the segmented uterine neck cell image is subjected to feature extraction operation and then subjected to dimension reduction operation via a genetic algorithm; in step S40, three single classifiers are integrated via fuzzy integration, and then the uterine neck cell image with reduced dimensions is subjected to identifying operation. The uterine neck cell image identification method provided in the invention can be used for precisely segmenting the cytoplasm and the cell nucleuses of uterine neck cells, segmenting speed can be increased, defects of low precision of one single classifier can be compensated by multiple classifier integration, and therefore uterine neck cell identification rate can be improved.

The invention discloses a high-resolution remote sensing image classifying method based on fusion of multiple classifiers. The high-resolution remote sensing image classifying method comprises the following steps that first, a training sample set is selected in an area of interest; second, the multiple classifiers are used for classifying remote sensing images; then, areas with the ground feature category classifying precision being lower than a threshold value a are classified again by using a voting method based on priori knowledge; at last, areas with the ground feature category classifying precision being lower than a threshold value b are classified by using a fuzzy decision template method, and finally the classified result of the target images is obtained. According to the high-resolution remote sensing image classifying method based on fusion of the multiple classifiers, the advantages of a single classifier are concentrated furthest, the disadvantages of the single classifier are restrained, the influences of 'same object with different spectrums ' and 'different objects with the same spectrum' on the classifying precision are lowered, and the precision of high-resolution remote sensing image classifying is improved.

The invention discloses an unbalanced data distribution-based multi-heterogeneous base classifier fusion classification method, and relates to an unbalanced data classification technology in the field of data mining. The method comprises the following steps of: preprocessing a sample by using a difference sampling rate-based resampling algorithm, including an oversampling and an under-sampling process, thereby distributing different samples to be classified for different base classifiers; calculating a classification error rate of each base classifier and further calculating the corresponding weight; counting respective results by an oversampling expert and an under-sampling expert; and fusing the final prediction result according to a classification strategy function to obtain the category of the sample. By using the multi-heterogeneous base classifier fusion classification method, important characteristics of a few types of samples are found in mass data, and the accuracy of the few types of samples can be effectively improved, so that the aim of improving the integral classification accuracy of a data set is fulfilled.

The invention relates to an energy efficiency evaluation method based on a multi-model fusion strategy. The problems that existing energy efficiency computing features are difficult to select, and the model evaluation result is inaccurate are solved. The method comprises the steps that 1, data is subjected to normalization processing, and a normalization training set is obtained; 2, the normalization training set obtained in the step 1 is subjected to feature selection, features are selected by adopting a fusion method of combining information gain with kernel principal component analysis, that is to say, after feature ordering is obtained through information gain calculation, verifying calculation is conducted through a principal component analysis method; 3, a multi-classifier fused evaluation module is built according to the step 1 and the step 2, a classification result obtained in the step 3 is subjected to cluster analysis, and the final cluster result is obtained. The energy efficiency evaluation method is applied to the field of energy efficiency effective evaluation.

The present invention discloses a credit evaluation method based on a fusion model. The method comprises: collecting individual personal credit data as a sample and simultaneously marking the credit rating; dividing the credit data into a plurality of training sets of equal element numbers through random sampling, and putting the training sets into different single classifiers, wherein each singleclassifier implements a classification algorithm; fusing results generated by each single classifier by using a fusion algorithm, extracting an optimal classification scheme, and recording the schemeby using a mathematical model to generate a preliminary model; and finally re-inputting the data to the preliminary model and verifying the data. The present invention also discloses an electronic device and a computer-readable storage medium applying to the method. According to the technical scheme of the present invention, multiple single classifiers are integrated to select the most appropriate classification scheme by means of an ensemble learning method, and respective weaknesses are overcome to exerting the greatest effect, so that the accuracy of the credit rating evaluation by using the fusion model can be improved.

The invention provides a subspace fusion-based protein-vitamin binding location point predicting method. The method comprises the steps of feature extracting and feature combining: extracting evolution information, secondary-level structural information the binding tendency information of a protein by respectively using a PSI-BLAST, a PSIPRED and a protein-vitamin binding location point tendency table, and converting amino acid residues in a protein sequence into a vector presentation mode by a sliding window and serial combination; using a multi-feature selection algorithm to perform feature selection on an original feature space for multiple times; forming a feature subspace by feature subsets obtained from feature selection every time and establishing multiple feature subspaces; training one SVM (support vector machine) classifier for each obtained feature subspace; fusing multiple SVM classifiers which are trained by a weighted average classifier fusing mode; performing protein-vitamin binding location prediction on the protein to be predicted based on a fused SVM predictor. The prediction method is fast in prediction sped and high in prediction precision.

The invention provides an automatic fruit grading system based on fusion of multiple classifiers. The automatic fruit grading system comprises a plurality of image capturing units, a feature extracting unit, a plurality of fruit classifying units and a fusing unit, wherein the image capturing units are respectively used for capturing image information of a same fruit; the feature extracting unit is used for analyzing the image information provided by the image capturing units so as to obtain the feature information of the fruit; each of the fruit classifying units is used for classifying the fruit by adopting a respective predetermined neural network classifier according to the obtained feature information of the fruit; and the fusing unit is used for fusing the classifying results of the fruit classifying units on the basis of a predetermined fusion strategy so as to obtain a grade result of the fruit. The automatic fruit grading system has a simple and stable structure and can be applied to more industrial fields of agricultural fruit production.

The present invention provides a color picture detecting method, characterized in that the input image is mapped to multiple different color spaces; the color images of the color space are processed edges detection respectively, obtaining the edge detecting result of each color space; the edge detecting results are syncretized. The invention processes edge detection in multiple color spaces, and uses class selector fusion technology integrated the detecting results, so as to combine the advantages of all kinds of color spaces expressed on the images, substantially utilizing the edge detecting results of all kinds of color space and improving the accuracy of the edge detection.

The invention discloses a method for diagnosing fusion faults of multiple classifiers on the basis of a fault type classification capacity evaluation matrix. On the basis of the measurement level output modes of the classifiers, the method for evaluating the classification capacity of the classifiers for multiple fault types on the basis of entropies of output results of the classifiers is provided, the evaluation matrix is obtained through calculation, a multi-classifier fusion basic model based on the fuzzy comprehensive evaluation method is constructed, the strategy-level fusion is conducted, and the final diagnosis is obtained.

The present invention discloses a physiological signal emotion identification method based on subjective and objective fusion of multiple classifiers. The method comprises that: a user experiences theuse of the product, and fills in the Chinese version of the PAD emotion inventory questionnaire; the heart rate and skin electrical signals of the user are collected during the product experience process, processing and feature extraction are carried out on the two kinds of objective physiological signals; the extracted heart rate features and skin electrical features are respectively trained andidentified by using SVM classifiers; and the identification results of each classifier are expressed in the form of probability of the target category, and the identification results are normalized;weight assignment is carried out on each classifier, and the particle swarm optimization algorithm is used to optimize the weights; and finally, identification results for different emotional categories are fused, and the type of emotion with the highest identification rate is taken as the final emotional state. According to the method disclosed by the present invention, the method of multi-classifier fusion is used to balance the decision results of subjective, objective and different physiological signals, so that the final identification result is more accurate and reliable.

The invention discloses a classifier fusion and diagnosis rule based PVC identification system and method. The system comprises a classification unit, a fusion unit and a discrimination unit; the classification unit comprises an LCNN classification module and an RNN classification module which process electrocardiogram data independently, the LCNN classification module includes m first classifiersof different structures, and outputs m first classification results at least, and the RNN classification module includes n second classifiers of different structures, and outputs n second classification results at least; the fusion unit carries out fusion decision on the first and second classification results according to a fusion decision rule to obtain a fusion result; and the discrimination unit discriminates non PVC data from PVC data of the fusion unit according to PVC pathologic characteristics, and obtains a PVC identification result. The classification results of the LCNN and RNN classifiers are fused, the PVC pathologic characteristics are combined, a machine learning and disease diagnosis rule combined method is used, and the integral classification performance and accuracy ofPVC identification are improved.

The invention discloses a plant drought stress diagnostic method and a device based on chlorophyll fluorescence imaging technology. The plant drought stress diagnostic method comprises following steps: 1, sample set plants with known drought stress diagnosis results are subjected to dark adaptation, canopy chlorophyll fluorescence image data of the sample set plants are collected, and related characteristic parameters are extracted; 2, based on the collected chlorophyll fluorescence image data and chlorophyll fluorescence image characteristic parameters, multiple classifiers fusion is adopted to establish a plant drought stress identification model; 3, chlorophyll fluorescence image data and characteristic parameters of plants to be detected are collected via the method disclosed in step 1, and are introduced into the plant drought stress identification model for drought stress diagnosis. The plant drought stress diagnostic method is capable of providing chlorophyll fluorescence parameters, realizing visualization of distribution of the fluorescence parameters in leaf space, displaying the heterogeneity of photosynthesis on leaf surfaces and among leaves, and realizing real-time drought stress diagnosis at high speed at high precision.

A multi-model fusion evaluation system, the invention relates to a multi-model fusion evaluation system. The present invention solves the problems of difficult selection of existing energy efficiency calculation features and inaccurate model evaluation results. The steps of the present invention are: Step 1: normalize the data to obtain a normalized training set; Step 2: perform feature selection on the normalized training set obtained in Step 1; The combined fusion method selects features; that is, after the feature ranking is obtained by information gain calculation, the principal component analysis method is used to do the check calculation. Step 3: Establish a multi-classifier fusion evaluation model based on Step 1 and Step 2 to obtain the classification results of energy efficiency evaluation; Step 4: Perform cluster analysis on the classification results obtained in Step 3 to obtain the final cluster results. The invention is applied in the field of effective evaluation of energy efficiency.

The invention relates to the computer image processing field and discloses an object detection system based on a multi-classifier, which comprises: a classifier training unit used to obtain N classifiers according to the training of a training set, wherein, N is more than one; a classifier selection unit used to select P classifiers from the N classifiers according to the computational amount and the classification performance to fusingly obtain a classifier set, wherein, P is more than one, and is less than or equal to N; a classifier distribution unit used to distribute the P classifiers to a plurality of different computing resources to respectively detect an unknown image to obtain P classifier results; a detection result fusion unit used to fuse the P classifier results to obtain an object detection result. The invention also provides a corresponding method. The invention aims at a plurality of different features in the image to respectively train a plurality of classifiers, and selects the classifiers suitable for distributed operation to be distributed to different computing resources to respectively detect the image, thereby improving the speed of object detection.