479 results about "Adaboost algorithm" patented technology

The invention discloses a method and a system for authenticating a shielded face, wherein the method comprises the following steps: S1) collecting a face video image; S2) preprocessing the collected face video image; S3) performing detection calculation on the shielded face, evaluating a position of a face image by utilizing a three-frame difference method according to motion information of a video sequence, and further confirming the position of the face according to an Adaboost algorithm; and S4) performing authenticating calculation on the shielded face, dividing a face sample into a plurality of sub-blocks, performing shielding distinguishment on the sub-blocks of the face by adopting a SVM(Support Vector Machine) binary algorithm combined with a supervising 1-NN k-Nearest neighbor method, if the sub-blocks are shielded, directly abandoning the sub-blocks, and if the sub-blocks are not shielded, extracting a corresponding LBP (Length Between Perpendiculars) textural feature vector for performing weighting identification, and then using a classifier based on a rectangular projection method to reduce feature matching times. According to the method for authenticating the shielded face, the detection rate and the detection speed for the local shielded face are effectively increased.

This invention provides an identification method and a device for countenance based on the video, which applies the ASM profile pick-up algorithm in the pick-up of character vectors and picks up the man-face image based on the position of the eyes to generate a normalized character face from the chin and picks up the most effective character in the character face to identify the countenance, which can eliminate the influence of illumination to make the right and left gray value almost the same with the variance.

The invention discloses a method for quickly and accurately detecting and tracking a human face based on a video sequence, which relates to the technical field of mode identification. The method comprises the following steps of: 1, extracting a video frame image from a video stream; 2, preprocessing the video frame image, namely compensating light rays, extracting skin color areas, performing morphological processing and combining the areas; 3, detecting the human face, namely representing the human face by using Harr-like characteristics and detecting the human face by using a cascaded Adaboost algorithm with an assistant decision function; 4, establishing the characteristics of the human face, namely detecting the area characteristics of the detected human face and the shape characteristics of the edge profile of the human face; 5, tracking the human face, particularly tracking the human face by using a human face area characteristic model when an intersection does not occur in a human face area, and further matching when the intersection occurs according to the shape characteristics of the edge profile of the human face; and 6, extracting the sequence of a human face image. By the technical method, the human face can be detected and tracked quickly and accurately on the basis of the video sequence.

The invention discloses a human face recognition method which is based on Gabor information fusion of visible light and near-infrared light. The method comprises: human face images under a visible light source and a near-infrared light source are respectively collected, Gabor features of the two images are respectively extracted to be fused at a feature layer; an AdaBoost algorithm is adopted to carry out the feature selection on the feature after the fusion, and the nearest neighbor classification is adopted to carry out the calculation and the classification on the similarity thereof. The human face recognition method has very high accuracy rate and excellent robustness of the impacts of light illumination on the human face recognition; in addition, compared with other methods, the human face recognition method has the advantages of small number of the used features, rapid classification speed, etc.

The invention provides a face detecting and tracking method and a device. The method comprises the steps of inputting a face image or a face video, preprocessing the face image or the face video in an illumination mode, detecting a face by usage of an Ada Boost algorithm, confirming an initial position of the face, and tracking the face by the usage of a Mean Shift algorithm. According to the face detecting and tracking method and the device, a self-adaptation local contrast enhancement method is provided to enhance image detail information in the period of image preprocessing, in order to increase robustness under different illumination conditions, face front samples under different illumination are added to training samples and accuracy of the face detection is increased by adoption of the Ada Boost algorithm in the period of face detection, in order to overcome the defect that using color of the Mean Shift algorithm is single, grads features and local binary pattern length between perpendiculars (LBP) vein features are integrated by adoption of the Mean Shift tracking algorithm in the period of face tracking, wherein the LBP vein features further considers using LBP local variance for expressing change of image contrast information, and accuracy of the face detection and the face tracking is improved.

The invention discloses a short-term and medium- and long-term electric power load prediction method based on machine learning model. Firstly, preprocessing is conducted on data, including smootheningabnormal data and filling missing data. Factors of affecting load changes will be analyzed, including historical data, time periodicity, and weather variable characteristics. Domestication will be conducted on all input variables for accelerating learning speed and raising prediction precision. The invention is advantageous in that linear regression is compared, and the performance of the vectorregression and gradient lifting regression in the short-term and medium- and long-term electric power load prediction is supported; with the prolongation of the prediction time, the performance of thegradient lifting regression model is better that that of the other two models; the AdaBoost algorithm which uses the gradient lifting tree as a basic classifier is brought forward, and load prediction is conducted, and the precision of electric power load prediction can be effectively raised.

The invention discloses a method for detecting train driver behavior and fatigue state on line and a detection system thereof. The method comprises: adopting a camera to shoot a train driver; carrying out human face detection on pictures collected by the camera; judging whether the driver does not perform duty or nods off; then, using an eye open image detector trained by the AdaBoost algorithm to find an eye open image from the human face image so as to judge whether the driver opens eyes or not; and calculating the ratio of eye closing time to total testing time in real time, namely the PERCLOS value, for marking fatigue state. The system comprises a camera, an embedded host, an alarming device, a user press key and a memorizer, wherein, the camera, the alarming device, the user press key and the storage are respectively connected with the embedded host. The invention fills the gap of the field of monitoring train driver behavior and detecting fatigue state and can monitor train driver behavior and detect fatigue state under train environment.

The invention provides an eye precision orientation technology in the field of human face identifying. It is characterized in that it provides an eye precision orientation method on complex background image. It adopts a micro structure character with high efficiency and high redundant to express eye mode local and area grey distributing character and uses AdaBoost computing method to choose the most differential micro structure character to form the strong classer.

The invention relates to a robust ear detection method which improves an AdaBoost arithmetic and belongs to the technical field of image mode identifying. The invention is characterized by proposing an ear detection method with excellent performances under a complex underground. The invention proposes four anisomerous Haar-like corner characteristics which are used for describing the grayscale changes on the partial areas of the ears; a policy of subsection selection is adopted for selecting the best sorting threshold of the Haar-like characteristics, thus reducing the sample training time; the weight of a weak sorter is modified for reducing the mistaken detection rate of the sorter; the threshold HW is set and eliminated according to the distribution change of the sample weight in the training, thereby preventing an over-studying phenomenon from being generated and leading the miss-detection rate and the mistaken direction rate of the ear detection to be reduced; besides, the invention also provides a single-ear detection policy for leading both the defection efficiency and the detection effect to be improved. The excellent performances of the robust ear detection method are shown on a PC machine and a DSP.

The invention discloses a generation device for a text classification model. The generation device comprises a memory and a processor. A model generation program capable of being operated on the processor is stored on the memory. When the program is performed by the processor, the program comprises the following steps of obtaining a word segmentation dictionary in the financial field and a text corpus in the financial field; selecting candidate new words from the text corpus and adding the candidate new words to the word segmentation dictionary; obtaining a sample set, and carrying out type labeling on training samples in the sample set; on the basis of the word segmentation dictionary in which the candidate new words are added, carrying out word segmentation on the training samples in thesample set through utilization of a preset word segmentation algorithm; and extracting word vectors, and on the basis of an adaboost algorithm, inputting the word vectors and labeled type informationinto a plurality of weak classifiers for training, thereby obtaining a text classification model. The invention also provides a generation method for the text classification model and a computer readable memory medium. According to the device, the method and the computer readable memory medium, the problem that in the prior art, emotion tendency classification cannot be carried out on the financial field texts is solved.

The invention discloses a pedestrian detection method based on video monitoring, which comprises the following steps of: rapidly detecting a pedestrian by utilizing an expanded gradient histogram characteristic and an Adaboost algorithm, and then further identifying and verifying the detected pedestrian by utilizing the gradient histogram characteristic and a support vector machine. In the invention, experiments based on a standard pedestrian database indicate that the method not only can greatly reduce the detection time, but also markedly improves the detection rate. As one of key technologies of target detection, pedestrian detection has broad application prospect in the fields of video monitoring safety, intelligent traffic management and the like.

In camera-based object labeling, boost classifier ƒT(x)=Σr=1Mβrhr(x) is trained to classify an image represented by feature vector x using a target domain training set DT of labeled feature vectors representing images acquired by the same camera and a plurality of source domain training sets DS<sub2>1< / sub2>, . . . , DS<sub2>N < / sub2>acquired by other cameras. The training applies an adaptive boosting (AdaBoost) algorithm to generate base classifiers hr(x) and weights βr. The rth iteration of the AdaBoost algorithm trains candidate base classifiers hrk(x) each trained on a training set DT∪DS<sub2>k< / sub2>, and selects hr(x) from previously trained candidate base classifiers. The target domain training set DT may be expanded based on a prior estimate of the labels distribution for the target domain. The object labeling system may be a vehicle identification system, a machine vision article inspection system, or so forth.

The invention relates to the field of the computer visual technology, in particular to a fast 3D skeleton model detecting method based on a depth camera. The fast 3D skeleton model detecting method based on the depth camera comprises the steps that a whole human body is shot by using the depth camera, human face detection is carried out in the image by using an Adaboost algorithm, and thus depth information of the human face is obtained; the body silhouette is extracted based on the depth information of the human face; detection verification is carried out on the detected body silhouette through a 'convex template' verification algorithm; after the verification succeeds, image smoothness processing is carried out on the body silhouette, and the skeleton line of the body silhouette is obtained through a detailing algorithm; characteristic points on the skeleton line of the body silhouette are extracted, the number and the positions of the characteristic points are corrected, and interference points are removed; the corrected characteristic points are verified, and accurate joint points and other characteristics are obtained by adopting a fast joint point extracting algorithm if the verification succeeds. The fast 3D skeleton model detecting method based on the depth camera is high in operating speed, low in computing complexity and adaptive to various complex backgrounds, and each frame of image only needs 5ms.

The present invention discloses a serial-parallel combined multi-mode emotion information fusion and identification method belonging to the emotion identification technology field. The method mainly comprises obtaining an emotion signal; pre-processing the emotion signal; extracting an emotion characteristic parameter; and fusing and identifying the characteristic parameter. According to the present invention, firstly, the extracted voice signal and facial expression signal characteristic parameters are fused to obtain a serial characteristic vector set, then M parallel training sample sets are obtained by the sampling with putback, and sub-classifiers are obtained by the Adabost algorithm training, and then difference of every two classifiers is measured by a dual error difference selection strategy, and finally, vote is carried out by utilizing the majority vote principle, thereby obtaining a final identification result, and identifying the five human basic emotions of pleasure, anger, surprise, sadness and fear. The method completely gives play to the advantage of the decision-making level fusion and the characteristic level fusion, and enables the fusion process of the whole emotion information to be closer to the human emotion identification, thereby improving the emotion identification accuracy.

In a method for detecting facial expressions of a portrait photo by an image capturing electronic device, a face captured in the portrait photo is detected. The position and range of the opened and closed facial features are detected, and the facial features within an identified range are magnified according to a specific proportion. A patch of facial features and their surroundings within a specific range is cut according to the magnified identified range, so that the patch can show a change of facial expressions and a specific range of their surroundings. A facial feature classifier is trained by a specific number of opened and closed facial feature samples based on the Adaboost algorithm and used for detecting the facial features in the patch to determine whether the facial feature is situated at an opened state or a closed state.

Disclosed is a high-speed train driver alertness detecting method based on face image and eye movement analysis. The method includes steps of acquiring images of the head of a driver by a primary camera and a secondary camera simultaneously, extracting characteristics of the head of the driver step by step by the aid of wavelet multi-scale, recognizing the head of the driver via a neural network training method, and detecting the face of the driver via an AdaBoost arithmetic; respectively starting a Harr characteristic and a two-dimensional orthogonal Log-Gabor filtering phase characteristic to detect human eyes under different illuminations; building an adaptive non-linear level-log model of the driver; realizing eye movement tracking analysis for the driver via a strong tracking particle filter arithmetic; and finally realizing weight sum for an eye movement characteristic fatigue factor value and six face images including PERCLOS, pupilla, blinking, nodding, yawning and face inclination so as to obtain an alertness value of the driver. The calculated alertness value is high in accuracy and robustness.

The invention provides a method for tracking a moving object. The method comprises the following steps of 1, accurately detecting the moving object by a rapid moving object detection method based on a codebook model; 2, initializing a weak classifier group of an online Adaboost algorithm to obtain a strong classifier, fusing local direction histogram characteristics and color characteristic in selection of characteristics of the moving object; 3, computing a characteristic matrix of an online Adaboost tracking algorithm and the weak classifier to obtain a confidence chart, updating the weak classifier according to the obtained moving object position by a CamShift tracking algorithm on the confidence chart, and finally obtaining a tracking result of a whole segment of video sequence. The method for tracking the moving object capable of effectively tracking the moving object which largely deforms within a short period of time, generates shelter even large-area shelter, is similar to the colors of the other moving objects in background, rapidly changes and accelerates is provided by the invention.

The invention discloses a hyper-spectral remote sensing image classifying method based on AdaBoost. A traditional mode identification method cannot meet the requirements of carrying out high-efficiency and high-precision classification on hyper-spectral data with high data dimensions and great data quantity; and although a neural network and a support vector machine can effectively classify remote sensing data, an ideal selection method of parameters does not exist. The hyper-spectral remote sensing image classifying method based on the AdaBoost comprises the following steps of: pre-processing the hyper-spectral data to remove abnormal wave bands influenced by factors including atmosphere absorption and the like; then utilizing MNF (Minimum Noise Fraction) conversion to carry out wave band preferential selection to achieve the aims of optimizing data, removing noises and reducing dimensions of the data; then, dividing a training sample and a test sample; selecting a decision stump as a weak classifier and utilizing an AdaBoost algorithm to train the weak classifier to obtain a strong classifier; selecting suitable iterations; and finally, utilizing a one-to-one method to establish a plurality of the classifiers. According to the hyper-spectral remote sensing image classifying method based on the AdaBoost, the convergence rate is enhanced and the classification performance of a hyper-spectral image is improved.

The invention discloses a work method of an identification and early-warning system based on pedestrians and bicycle riders in front of a vehicle. An offline training module selects positive and negative samples of the upper half of the body from a video shot by the practical vehicle, an Adboost algorithm is used for training to obtain a cascaded classifier for identifying the upper half of the body, the cascaded classifier is provided for an online detection module, the online detection module uses a CCD camera to collect images, a video collection card is used for conversion, and an improved multi-scale scanning method is used for preprocessed images to obtain a subwindow; and the online detection module selects an identified target frame, and transmits the target frame to a collision early-warning module, and the collision early-warning module uses a monocular visual range-finding geometric model to calculate the horizontal distance X and the vertical distance D between a target and the vehicle, the vertical speed Vy of the target and the vertical collision time TTC; and such information and the speed of the vehicle u are integrated to determine the danger degree of the target, a driver is prompted timely, and accidents can be reduced o protect the pedestrians and bicycle riders.

The invention relates to the technical field of pattern identification and artificial intelligence, in particular relates to an extracting method for smiling face identification on a picture of a human face. The invention provides the extracting method for smiling face identification on the picture of the human face, wherein smiling faces and non-smiling faces can be differentiated by using characteristics of a layered gradient column diagram PHOG (PYRAMID HISTOGRAM OF ORIENTATIONGRADIENTS). The smiling face identification method comprises the following steps: (11) human face detection based on Haar characteristics and AdaBoost algorithm; (12) location of a mouth area of the human face; (13) characteristic extracting algorithm of the layered gradient column diagram; (14) training and identification of a support vector machine (SVM). The method differentiates smiling faces and non-smiling faces by proposing to use PHOG characteristics against over-high Gabor characteristic dimension, and can acquire experimental results equivalent to the identification rate of the Gabor characteristics by lower characteristic dimension. Therefore, the method can greatly improve the efficiency of algorithm, and is easier for usage in reality.

In a method for detecting facial expressions of a portrait photo by an image capturing electronic device, a face captured in the portrait photo is detected. The position and range of the opened and closed facial features are detected, and the facial features within an identified range are magnified according to a specific proportion. A patch of facial features and their surroundings within a specific range is cut according to the magnified identified range, so that the patch can show a change of facial expressions and a specific range of their surroundings. A facial feature classifier is trained by a specific number of opened and closed facial feature samples based on the Adaboost algorithm and used for detecting the facial features in the patch to determine whether the facial feature is situated at an opened state or a closed state.

The invention discloses a grounding grid corrosion rate level prediction method which comprises the following steps: (1) inputting training sample data; (2) randomly sampling training samples according to a bootstrap sampling principle in a Bagging algorithm, forming training sample bootstrap subsets with the number of M, and constituting training sample bootstrap subset data sets; (3) structuring a weak classifier model according to a k-nearest neighbor (KNN) algorithm, sequentially training the training sample bootstrap subsets with the number of M, and obtaining weak classifiers with the number of M; (4) structuring a strong classifier model according to an Adaboost algorithm; (5) inputting to-be-tested sample data, predicting a grounding grid corrosion rate level, obtaining a predicting result, and displaying the predicting result through a displayer. The grounding grid corrosion rate level prediction method has the advantages of being novel and reasonable in design, convenient and fast to use and operate, high in predicting precision, capable of achieving an accurate prediction to the grounding grid corrosion rate level by means of a small amount of data samples which are measured in the prior art, low in implementation cost, strong in practicability and high in value of popularization and application.

The invention discloses a method for checking class attendance on the basis of multi-face data acquisition strategies and deep learning. By the aid of the method, the technical problem of low recognition rates of existing methods for checking attendance on the basis of face recognition can be solved. The technical scheme includes that multiple objects are detected and extracted by the aid of AdaBoost algorithms and skin color models. Only a piece of video needs to be shot on every face participating in attendance checking at one step, faces in video sequences are detected and extracted, and face databases can be completely created. The method has the advantages that learning can be carried out on face features in the face databases in different scenes by the aid of simplified LeNet-5 models on the basis of depth convolutional neural network LeNet-5 models by the aid of processes for recognizing the faces on the basis of deep learning, and novel features can be represented by means of multilayer nonlinear transformation; intra-class change of illumination, noise, attitude, expression and the like is removed from the novel features as much as possible, inter-class change generated by identity difference is reserved, and accordingly the face recognition rates of the processes for recognizing the faces in practical complicated scenes can be increased.

The present invention relates to a video-based method of intense passenger flow counting and automatic detection of pedestrian speed, belonging to the computer vision technical field. The present invention adopts video capture device and process algorithm. Wherein, the video image capture device captures images through completed circuit television (CCTV ), normally with a camera set over the top of entrance and exit of passage way for real-time capture of video images of passengers in and out. The processor adopts computer vision algorithm to process the captured video images and identify the faces with Adaboost algorithm, with Harr characteristics as input, weak classifier weight combination to form a strong classifier, the strong classifier to form a waterfall-like cascading. Then calculation is triggered. The walking speed calculation of pedestrians is mainly based on tracking of the faces. The tracking starts as the pedestrian faces enter the detection zone and stops as the faces exit the detection zone. The keys to the walking speed calculation of pedestrians are targeting and tracking.

The invention discloses a Haar and HoG characteristic based preceding car detection method. The method includes: 1), manually selecting a large number of vehicular pictures and non-vehicular pictures as positive samples and negative samples of a training set, and formatting the positive samples and the negative samples to be under 24X24 pixel; 2), using Haar characteristics and HoG characteristics to respectively characterize each of the positive samples and the negative samples to form characteristic vectors; 3) respectively establishing a weak classifier aiming at two characteristic vectors formed according to the Haar characteristics and the HoG characteristics; 4), utilizing the cascaded Adaboost algorithm to train the weak classifiers to obtain a cascaded vehicular strong classifier; and 5), inputting sub-images of different sizes and different positions into the cascaded vehicular strong classifier for judgment according to preceding road video images obtained by a vehicular camera, and judging the sub-images where a vehicle is positioned to be a preceding vehicle. The Haar and HoG characteristic based preceding car detection method is good in instantaneity and high in robustness, and has active influences on guaranteeing vehicular safe traveling, human and property safety.

The invention discloses a rapid pedestrian detection method based on computer vision. The method comprises the following steps: firstly, acquiring a pedestrian video image, then establishing a background model for a video image frame by a ViBe algorithm, and segmenting out a foreground area by the ViBe algorithm combined with a frame difference method; secondly, calculating the coordinates of the highest point of each target block outline, thereby finishing location of a head candidate area; and finally, traversing the head candidate area by a target detection window to obtain a sample to be detected, inputting the sample to be detected in a composite head detection trained by an AdaBoost algorithm, and judging whether the sample is a human head sample to obtain a pedestrian detection result. According to the method with the technical scheme, the technical problem of poor robustness and real-time performance in the prior art is effectively solved.

The invention provides a face authentication method and device, and belongs to the field of image processing and pattern recognition. The method comprises the following steps: extracting Gabor texture features of an acquired face image sample pair through a Gabor filter, and separating real parts from virtual parts of the extracted Gabor texture features to obtain a plurality of feature graphs; summarizing the plurality of obtained feature graphs through an LBP (Local Binary Pattern) statistical histogram to serve as face feature expression of the face image pair; performing feature selection on the face feature expression of the face image sample pair through an AdaBoost (Adaptive Boosting) algorithm; and performing feature dimension reduction, feature judgment processing and the like on the face feature expression of the face image sample pair subjected to the feature selection through a PCA (Principal Components Analysis) algorithm and an LDA (Linear Discriminant Analysis) algorithm in sequence. Compared with the prior art, the face authentication method provided by the invention has the advantages of full extraction of sample texture information, low sample quantity demand, short algorithm time and low space complexity.

The invention discloses a method of video monitor object automatic detection and a system thereof. Combined with object motion information and form information in a video, based on a Gentle AdaBoost algorithm and an expanded Harr characteristic, a classifier is trained and object detection in the video is carried out automatically, a return value is determined after employing a training sample window to pass through each layer of a cascade classifier, when the return value is a positive number, an object is searched in real time in video monitoring and is highlighted, and a problem of low inquire efficiency of mass video data in the prior art is solved. The method and the system have the characteristics of simple design, fast detection speed, high precision and strong robustness, efficiency of extracting a characteristic in the video is raised, and the method and the system can be widely used for pedestrian detection and tracking in the video retrieval field.