37 results about "Score fusion" patented technology

The voice print system of the present invention is a subword-based, text-dependent automatic speaker verification system that embodies the capability of user-selectable passwords with no constraints on the choice of vocabulary words or the language. Automatic blind speech segmentation allows speech to be segmented into subword units without any linguistic knowledge of the password. Subword modeling is performed using a multiple classifiers. The system also takes advantage of such concepts as multiple classifier fusion and data resampling to successfully boost the performance. Key word / key phrase spotting is used to optimally locate the password phrase. Numerous adaptation techniques increase the flexibility of the base system, and include: channel adaptation, fusion adaptation, model adaptation and threshold adaptation.

Provided are an object verification apparatus and method. The object verification apparatus includes a matching unit performing a plurality of different matching algorithms on a query image and generating a plurality of scores; a score normalization unit normalizing each of the generated scores to be adaptive to the query image; a weight estimation unit estimating weights of the normalized scores based on the respective matching algorithms applied; and a score fusion unit fusing the normalized scores by respectively applying the weights estimated by the weight estimation unit to the normalized scores.

The invention discloses a face recognition method based on combining face attribute information, and belongs to the technical field of digital image processing. Aiming at the technical problem that the existing fusion method needs to train multiple DCNN networks, and then performs score fusion or feature fusion for further training, the work task is heavy and complex, which is not conducive to practical application, and discloses a new fusion method of identity information and attribute information. way to improve the accuracy of face recognition. The invention fuses the face identity authentication network and the attribute recognition network to form a fusion network, and adopts the joint learning method to learn the identity feature and the face attribute feature at the same time, which not only improves the correct rate of face recognition, but also can predict the face. It is a multi-task network; it adopts a cost-sensitive weighting function, so that it does not depend on the data distribution of the target domain, and realizes balanced training in the source data domain; and the modified fusion framework only adds a small number of parameters, The additional computational load is relatively small.

The present invention relates to the field of computer vision technology, in particular to a face attribute classification system based on a two-way ladder structure, aiming to solve how to make full use of the features of different levels in the deep network, and the correspondence between features of different levels and different face attributes relationship to improve the accuracy of face attribute classification. For this purpose, the face attribute classification system based on the two-way Ladder structure provided by the present invention includes a two-way Ladder self-encoder module, an adaptive attention module and an adaptive scoring fusion module; the two-way Ladder self-encoder module includes an encoder module and a decoder The adaptive attention module includes a plurality of attention submodules; the adaptive score fusion module is configured to obtain the face attribute classification result of the face image to be tested according to the output result of the encoder module and the output result of the attention submodule . Based on the above structure, different levels of encoding features and decoding features can be fully utilized to improve the accuracy of face attribute classification.

The invention discloses a fingerprint score fusion system and a fingerprint score fusion method based on Sigmoid expansion. The system comprises a training process and a testing process; the training process is realized via a training image preprocessing unit, a training feature extracting unit, a training feature aligning unit and a Sigmoid parameter training unit; and the testing process is realized via a testing image preprocessing unit, a testing feature extracting unit, a testing feature aligning unit and a score fusion unit. According to the score fusion unit, Sigmoid expansion is conducted on each score by using parameters output from the Sigmoid parameter training unit respectively; the scores after the Sigmoid expansion are fused; and a result after the fusion is taken as a final matching score. The final matching score is used for deciding whether the input fingerprint feature and the template fingerprint feature are originated from the same finger, thereby finishing fingerprint matching.