检索结果-内蒙古大学图书馆

deep Feedforward Sequential Memory Networks Based Mispronunciation Detection for Tibetan Students' Mandarin 2

学校读者我要写书评

暂无评论

Deep Feedforward Sequential Memory Networks Based Mispronunc...

2nd International Conference on Information Science and Education, ICISE-IE 2021

作者： Gan, Zhenye Zhao, Tianqin Yu, Xinke Yang, Hongwu College of Physics and Electronic Engineering Northwest Normal University Engineering Research Center of Gansu Province for Intelligent Information Technology and Application LanZhou China School of Educational Technology Northwest Normal University National and Local Joint Engineering Laboratory for Learning Analytics Technology of Internet Education Data LanZhou China

ISBN: (纸本)9781665438292

Computer assisted pronunciation training system (CAPT) can detect the wrong pronunciation produced by nonnative speakers and provide positive feedback. CAPT is helpful to improve the pronunciation level for L2 learners' accurately. Tibetan students' Mandarin is influenced by their native language pronunciation habits. So, this feature leads to their pronunciation being obviously different from that of standard mandarin. This paper used CNN model as a fundamental, and we introduced acoustic model: DFSMN and CTC. This acoustic model implemented a method of speech recognition on Tibetan students' mandarin mispronunciation detection. In order to continue improving the detection performance, we used extended initial final (XIF) as bias primitives and design 64 bias types. Experiment results show that the proposed method in this paper can effectively detect mispronunciation and provide correct feedback, with the DA of 88.02%, FRR of 7.95% and FAR of 25.74%. © 2021 IEEE

关键词： Students

Every Pixel Counts: Unsupervised geometry learning with holistic 3d motion understanding

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Yang, Zhenheng Wang, Peng Wang, Yang Xu, Wei Nevatia, Ram University of Southern California Baidu Research National Engineering Laboratory for Deep Learning Technology and Applications

learning to estimate 3D geometry in a single image by watching unlabeled videos via deep convolutional network has made significant process recently. Current state-of-the-art (SOTA) methods, are based on the learning framework of rigid structure-from-motion, where only 3D camera ego motion is modeled for geometry estimation. However, moving objects also exist in many videos, e.g. moving cars in a street scene. In this paper, we tackle such motion by additionally incorporating per-pixel 3D object motion into the learning framework, which provides holistic 3D scene flow understanding and helps single image geometry estimation. Specifically, given two consecutive frames from a video, we adopt a motion network to predict their relative 3D camera pose and a segmentation mask distinguishing moving objects and rigid background. An optical flow network is used to estimate dense 2D per-pixel correspondence. A single image depth network predicts depth maps for both images. The four types of information, i.e. 2D flow, camera pose, segment mask and depth maps, are integrated into a differentiable holistic 3D motion parser (HMP), where per-pixel 3D motion for rigid background and moving objects are recovered. We design various losses w.r.t. the two types of 3D motions for training the depth and motion networks, yielding further error reduction for estimated geometry. Finally, in order to solve the 3D motion confusion from monocular videos, we combine stereo images into joint training. Experiments on KITTI 2015 dataset show that our estimated geometry, 3D motion and moving object masks, not only are constrained to be consistent, but also significantly outperforms other SOTA algorithms, demonstrating the benefits of our approach. Copyright © 2018, The Authors. All rights reserved.

关键词： Pixels

Occlusion aware unsupervised learning of optical flow

学校读者我要写书评

暂无评论

arXiv 2017年

作者： Wang, Yang Yang, Yi Yang, Zhenheng Zhao, Liang Wang, Peng Xu, Wei Baidu Research University of Southern California National Engineering Laboratory for Deep Learning Technology and Applications

It has been recently shown that a convolutional neural network can learn optical flow estimation with unsuper-vised learning. However, the performance of the unsuper-vised methods still has a relatively large gap compared to its supervised counterpart. Occlusion and large motion are some of the major factors that limit the current unsuper-vised learning of optical flow methods. In this work we introduce a new method which models occlusion explicitly and a new warping way that facilitates the learning of large motion. Our method shows promising results on Flying Chairs, MPI-Sintel and KITTI benchmark datasets. Espe-cially on KITTI dataset where abundant unlabeled samples exist, our unsupervised method outperforms its counterpart trained with supervised learning. Copyright © 2017, The Authors. All rights reserved.

关键词： Optical flows

TPOS Tagging Method Based on BiLSTM_CRF Model 5th

学校读者我要写书评

暂无评论

TPOS Tagging Method Based on BiLSTM_CRF Model

5th International Conference of Pioneer Computer Scientists, Engineers and Educators, ICPCSEE 2019

作者： Wang, Lili Chen, Ziyan Yang, Hongwu College of Physics and Electronic Engineering Northwest Normal University Lanzhou730070 China Engineering Research Center of Gansu Province for Intelligent Information Technology and Application Lanzhou730070 China National and Provincial Joint Engineering Laboratory of Learning Analysis Technology in Online Education Lanzhou730070 China

ISBN: (纸本)9789811501173

Part of speech (POS) tagging determines the attributes of each word, and it is the fundamental work in machine translation, speech recognition, information retrieval and other fields. For Tibetan part-of-speech (TPOS) tagging, a tagging method is proposed based on bidirectional long short-term memory with conditional random field model (BiLSTM_CRF). Firstly, the designed TOPS tagging set and manual tagging corpus were used to get word vectors by embedding Tibetan words and corresponding TPOS tags in continuous bag-of-words (CBOW) model. Secondly, the word vectors were input into the BiLSTM_CRF model. To obtain the predictive score matrix, this model using the past input features and future input feature information respectively learned by forward long short-term memory (LSTM) and backward LSTM performs non-linear operations on the softmax layer. The prediction score matrix was input into the CRF model to judge the threshold value and calculate the sequence score error. Lastly, a Tibetan part of speech tagging model was got based on the BiLSTM_CRF model. The experimental results indicate that the accuracy of TPOS tagging model based on the BiLSTM_CRF model can reach 92.7%. © Springer Nature Singapore Pte Ltd. 2019.

关键词： Long short-term memory

FusionPainting: Multimodal fusion with adaptive attention for 3D object detection

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Xu, Shaoqing Zhou, Dingfu Fang, Jin Yin, Junbo Bin, Zhou Zhang, Liangjun Beihang University Beijing100083 China Robotics and Autonomous Driving Laboratory Baidu Research National Engineering Laboratory of Deep Learning Technology and Application China Beijing Institute of Technology Beijing China

Accurate detection of obstacles in 3D is an essential task for autonomous driving and intelligent transportation. In this work, we propose a general multimodal fusion framework FusionPainting to fuse the 2D RGB image and 3D point clouds at a semantic level for boosting the 3D object detection task. Especially, the FusionPainting framework consists of three main modules: a multi-modal semantic segmentation module, an adaptive attention-based semantic fusion module, and a 3D object detector. First, semantic information is obtained for 2D image and 3D Lidar point clouds based on 2D and 3D segmentation approaches. Then the segmentation results from different sensors are adaptively fused based on the proposed attention-based semantic fusion module. Finally, the point clouds painted with the fused semantic label are sent to the 3D detector for obtaining the 3D objection results. The effectiveness of the proposed framework has been verified on the large-scale nuScenes detection benchmark by comparing with three different baselines. The experimental results show that the fusion strategy can significantly improve the detection performance compared to the methods using only point clouds, and the methods using point clouds only painted with 2D segmentation information. Furthermore, the proposed approach outperforms other state-of-the-art methods on the nuScenes testing benchmark. Code will be available at https://***/Shaoqing26/FusionPainting/. © 2021, CC BY-NC-ND.

关键词： Object recognition

LEGO: learning edge with geometry all at once by watching videos

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Yang, Zhenheng Wang, Peng Wang, Yang Xu, Wei Nevatia, Ram University of Southern California Baidu Research National Engineering Laboratory for Deep Learning Technology and Applications

learning to estimate 3D geometry in a single image by watching unlabeled videos via deep convolutional network is attracting significant attention. In this paper, we introduce a "3D as-smooth-as-possible (3D-ASAP)" prior inside the pipeline, which enables joint estimation of edges and 3D scene, yielding results with significant improvement in accuracy for fine detailed structures. Specifically, we define the 3D-ASAP prior by requiring that any two points recovered in 3D from an image should lie on an existing planar surface if no other cues provided. We design an unsupervised framework that Learns Edges and Geometry (depth, normal) all at Once (LEGO). The predicted edges are embedded into depth and surface normal smoothness terms, where pixels without edges in-between are constrained to satisfy the prior. In our framework, the predicted depths, normals and edges are forced to be consistent all the time. We conduct experiments on KITTI to evaluate our estimated geometry and CityScapes to perform edge evaluation. We show that in all of the tasks, ***, normal and edge, our algorithm vastly outperforms other state-of-the-art (SOTA) algorithms, demonstrating the benefits of our approach. Copyright © 2018, The Authors. All rights reserved.

关键词： Geometry

Coarse-to-Fine Cascaded Networks with Smooth Predicting for Video Facial Expression Recognition

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Xue, Fanglei Tan, Zichang Zhu, Yu Ma, Zhongsong Guo, Guodong University of Chinese Academy of Sciences Beijing China Key Laboratory of Space Utilization Technology and Engineering Center for Space Utilization Chinese Academy of Sciences Beijing China Institute of Deep Learning Baidu Research Beijing China National Engineering Laboratory for Deep Learning Technology and Application Beijing China

Facial expression recognition plays an important role in human-computer interaction. In this paper, we propose the Coarse-to-Fine Cascaded network with Smooth Predicting (CFC-SP) to improve the performance of facial expression recognition. CFC-SP contains two core components, namely Coarse-to-Fine Cascaded networks (CFC) and Smooth Predicting (SP). For CFC, it first groups several similar emotions to form a rough category, and then employs a network to conduct a coarse but accurate classification. Later, an additional network for these grouped emotions is further used to obtain fine-grained predictions. For SP, it improves the recognition capability of the model by capturing both universal and unique expression features. To be specific, the universal features denote the general characteristic of facial emotions within a period and the unique features denote the specific characteristic at this moment. Experiments on Aff-Wild2 show the effectiveness of the proposed CFSP. We achieved 3rd place in the Expression Classification Challenge of the 3rd Competition on Affective Behavior Analysis in-the-wild. The code will be released at https://***/BR-IDL/PaddleViT. © 2022, CC BY.

关键词： Forecasting

IAFA: Instance-aware feature aggregation for 3D object detection from a single image

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Zhou, Dingfu Song, Xibin Dai, Yuchao Yin, Junbo Lu, Feixiang Fang, Jin Liao, Miao Zhang, Liangjun Baidu Research China National Engineering Laboratory of Deep Learning Technology and Application Beijing China Northwestern Polytechnical University Xi’an China Beijing Institute of Technology Beijing China

3D object detection from a single image is an important task in Autonomous Driving (AD), where various approaches have been proposed. However, the task is intrinsically ambiguous and challenging as single image depth estimation is already an ill-posed problem. In this paper, we propose an instance-aware approach to aggregate useful information for improving the accuracy of 3D object detection with the following contributions. First, an instance-aware feature aggregation (IAFA) module is proposed to collect local and global features for 3D bounding boxes regression. Second, we empirically find that the spatial attention module can be well learned by taking coarse-level instance annotations as a supervision signal. The proposed module has significantly boosted the performance of the baseline method on both 3D detection and 2D bird-eye’s view of vehicle detection among all three categories. Third, our proposed method outperforms all single image-based approaches (even these methods trained with depth as auxiliary inputs) and achieves state-of-the-art 3D detection performance on the KITTI benchmark. © 2021, CC BY-NC-SA.

关键词： Object recognition

Targeted hydrolysis of α-lactalbumin to reduce allergenicity based on the synergistic effect of pepsin and papain

学校读者我要写书评

暂无评论

International Journal of Biological Macromolecules 2025年第Pt 4期311卷 144016页

作者： Zhao, Qing Mu, Guangqing Zheng, Yijia Du, Wenjiao Zhao, Anqi Sun, Qi Sun, Peng Wu, Xiaomeng Kong, Fanhua School of Food Science and Technology Dalian Polytechnic University Liaoning Dalian116034 China Dalian Key Laboratory of Functional Probiotics School of Food Science and Technology Dalian Polytechnic University Dalian116034 China SKL of Marine Food Processing & Safety Control National Engineering Research Center of Seafood Collaborative Innovation Center of Seafood Deep Processing National & Local Joint Engineering School of Food Science and Technology Dalian Polytechnic University Dalian116034 China Laboratory for Marine Bioactive Polysaccharide Development and Application Liaoning Key Laboratory of Food Nutrition and Health School of Food Science and Technology Dalian Polytechnic University Dalian116034 China Department of Agricultural Food and Nutritional Science University of Alberta EdmontonABT6G 2P5 Canada

Enzymolysis can effectively reduce the allergenicity of α-lactalbumin (ALA). Herein, complex enzymes targeted enzymolysis technology was developed to reduce the allergenicity of ALA, and the allergenicity of hydrolysates (WM) were investigated from multiple perspectives. The results showed that compared with the single enzyme hydrolysates, the degree of hydrolysis and antigen reduction rate of WM were increased by 5.57 % and 25.31 %, and linear allergenic epitopes were further eliminated. The spatial conformation of WM was disrupted, manifested by decreasing in α-helix content and surface hydrophobicity. In addition, WM could alleviate allergic symptoms in mice, reduced the binding ability with IgE and IgG1 and the release of mast cells degranulation, maintained the balance of Th1/Th2 cells and intestinal barrier function. Spearman's correlation analysis further revealed the correlation between the relative abundance of marker genus and allergy related indicators. This study will provide theoretical basis for the development of low allergenic dairy-based materials. © 2025 Elsevier B.V.

关键词： Allergens