452 results about "Multi feature fusion" patented technology

Uterine cervical cancer Computer-Aided-Diagnosis (CAD) according to this invention consists of a core processing system that automatically analyses data acquired from the uterine cervix and provides tissue and patient diagnosis, as well as adequacy of the examination. The data can include, but is not limited to, color still images or video, reflectance and fluorescence multi-spectral or hyper-spectral imagery, coherent optical tomography imagery, and impedance measurements, taken with and without the use of contrast agents like 3-5% acetic acid, Lugol's iodine, or 5-aminolevulinic acid. The core processing system is based on an open, modular, and feature-based architecture, designed for multi-data, multi-sensor, and multi-feature fusion. The core processing system can be embedded in different CAD system realizations. For example: A CAD system for cervical cancer screening could in a very simple version consist of a hand-held device that only acquires one digital RGB image of the uterine cervix after application of 3-5% acetic acid and provides automatically a patient diagnosis. A CAD system used as a colposcopy adjunct could provide all functions that are related to colposcopy and that can be provided by a computer, from automation of the clinical workflow to automated patient diagnosis and treatment recommendation.

A method for identity recognition based on multiple feature fusion for an eye image, which comprises steps of registering and recognizing, wherein the step of registering comprises: obtaining a normalized eye image and a normalized iris image, for a given registered eye image, and extracting a multimode feature of an eye image of a user to be registered, and storing the obtained multimode feature of the eye image as registration information in a registration database; and the step of recognizing comprises:obtaining a normalized eye image and a normalized iris image, for a given recognized eye image,extracting a multimode feature of an eye image of a user to be recognized,comparing the extracted multimode feature with the multimode feature stored in the database to obtain a matching score, and obtaining a fusion score by fusing matching scores at score level, and performing the multiple feature fusion identity recognition on the eye image by a classifier. The present invention recognizes identity by fusing multiple features of eye regions on a human face, and thus achieves high recognition accuracy and is suitable for applications of high security level.

The invention discloses a face-shielding detecting method based on multi-feature fusion, which is realized with the support of a digital camera and a digital signal processing chip. The method is characterized by comprising the following steps of: using the digital camera to acquire a digital video and converting the digital video into a digital image; obtaining a face image from the digital image by using a face detecting algorithm; aligning and zooming the face image and specifying the face image to a fixed resolution; then dividing the face image into a plurality of cells; and computing feature vectors of all cells; and circularly judging whether the face is shielded, wherein a shielding judgment rule is that the total number of the shielding cells is over a set threshold, or the number of the adjacent shielding cells is over a set threshold. A classifier is obtained by using a method of fusing a plurality of textural features and using an SVM method to train. The face-shielding detecting method based on multi-feature fusion has good classification performance and robustness, which can be widely applied to various monitoring occasions to judge whether a deliberate shielding behavior exists so as to screen out the suspicious personnel.

An intelligent house purchasing assistant based on a dialogue robot is provided. The invention discloses a multi-feature fusion unsupervised semantic matching algorithm model and an intelligent housepurchase dialogue robot system based on the model. The multi-feature fusion unsupervised semantic matching algorithm model considers the relevance and semantic relevance of statistical information, and the similarity calculation is carried out from both the statistical and semantic perspective. Finally, the similarity calculation results are more accurate and reliable, and the recall and matchingprocess of user problems and knowledge base problems can be quickly realized by using the fusion method to further improve the effect. The intelligent house purchasing dialogue robot system can realize the question-answer interaction, discard the existing house purchasing information acquisition mode of searching, save time, improve the user experience, and actively display the supporting facilities around the house property and the future trend of the house price for the user according to the user question-answer semantics, so that the user can further understand the house property information.

The invention discloses an infrared human body target recognition method based on multi-feature fusion and multi-core transfer learning. Firstly, based on the special imaging mechanism of infrared image, an improved CLBP feature is extracted from the target in the infrared scene, MSF-CLBP is used to express the texture information, and an improved local HOG feature-HOG-FV is used to express the shape information. In order to explore the effective features in the target; secondly, for the two heterogeneous features extracted above, the method of tandem fusion is used for feature fusion, so thatthe description of the target feature information is more objective and comprehensive. Finally, a classifier combining multi-core classification and TrAdaBoost migration learning framework is designed, which effectively solves the problem of insufficient image of labeled infrared samples, and enhances the distinguishability of the data to be classified to obtain better recognition results. The method starts from the improvement of feature extraction and the design of classifier, improves the expression of feature information, and improves the performance of infrared target recognition in complex background.

The invention provides a fault feature screening method based on weighted multi-feature fusion and SVM classification. The method comprises the following steps of 1, obtaining the time series data; 2,extracting a time domain (T), a frequency domain (P), the energy (E) and an entropy feature (S), and forming a high-dimensional feature set (Q); 3, screening out the features (Q1) with the diagnosisrate greater than 50%, carrying out the correlation analysis, and removing the features (Q2) with the similarity greater than 85%; 4, selecting a feature with the highest score through PCA and a loadscoring method to form a new sub-feature set (T3, P3, E3, S3); 5, carrying out SVM diagnosis on the T3, the P3, the E3 and the S3, and obtaining a weight Wi according to the diagnosis rate Ri; 6, performing the weighted fusion of the features; and 7, inputting the fused features into a classifier for diagnosis. Through the above steps, a group of optimal features capable of maintaining the fault intrinsic information is obtained, the original failure information represented by the features is ensured, the fault diagnosis accuracy is improved, and the method is of great significance to the efficient mechanical fault diagnosis.

The invention relates to an audio and video speech recognition technology, and provides a cross-modal multi-feature fusion audio and video speech recognition method and system in consideration of thesituation that speech interaction is easily affected by complex environmental noise, and facial motion information is acquired through a video and is relatively stable in an actual robot application environment. According to the method, speech information, visual information and visual motion information are fused through an attention mechanism, and the speech content expressed by a user is acquired more accurately by using the relevance among different modes, so that the speech recognition precision under the condition of complex background noise is improved, the speech recognition performance in human-computer interaction is improved, and the problem of low pure-speech recognition accuracy in a noise environment is effectively solved.

The invention discloses an encrypted malicious traffic detection method. According to the method, a Wreshark tool is utilized to process a traffic packet; filtering out invalid IP checksums, preprocessing the sample set and marking malicious / benign tags; performing preliminary feature extraction on the preprocessed traffic packet; constructing three feature subsets for the preliminarily extracted features, and standardizing and encoding the three feature subsets; carrying out feature dimension reduction on each type of feature subsets by adopting a machine learning or principal component analysis method; respectively establishing a random forest, an XGBoost classifier model and a Gaussian naive Bayes classifier model for the three feature subsets; the three classifier models are combined according to a Stacking strategy to form a DMMFC detection model; performing stream fingerprint fusion on the three feature subsets to form a sample set, dividing the sample set into a training set and a test set, and training a model; testing the model, and evaluating the test effect of the DMMFC model by using the evaluation indexes of the accuracy rate, the F1 score and the false alarm rate; encrypted malicious traffic detection is performed by adopting a method of combining multi-feature fusion and a Stacking strategy, and the method has relatively high detection capability.

The invention discloses a method and a device for detecting a human-object interaction relationship in a video, and the method comprises the steps: extracting a space-time track of a subject and an object in the video through target track detection, recognizing a human-object interaction relationship HOI through interaction joint reasoning based on a result of target track detection, and extracting multimode features from a target space-time track through interaction joint reasoning, performing joint reasoning by using a multi-feature fusion mode according to human and object fusion semantic features, human body behavior visual description features and human and object space-time relative motion features, and predicting human and object interaction actions on a video segment to obtain a predicted interaction category label, namely a human and object interaction relationship. The invention provides a time-space domain-based human-object interaction detection method, which successfully discovers, locates and classifies human-object pairs and interaction relationships between the human-object pairs in a video by capturing advanced semantic information of a scene from a complex video vision scene.