A5101775309- Advanced Neural Network Applications
- Domain Adaptation and Few-Shot Learning
- Cryptography and Data Security
- Video Surveillance and Tracking Methods
- Advanced Optical Network Technologies
- Multimodal Machine Learning Applications
- Image Processing Techniques and Applications
- Gastric Cancer Management and Outcomes
- Software-Defined Networks and 5G
- Colorectal Cancer Screening and Detection
- Privacy-Preserving Technologies in Data
- Advanced Image and Video Retrieval Techniques
- Advanced Steganography and Watermarking Techniques
- Advanced Photonic Communication Systems
- Machine Learning and ELM
- Big Data and Digital Economy
- Complexity and Algorithms in Graphs
- Physical Unclonable Functions (PUFs) and Hardware Security
- Teaching and Learning Programming
- 3D Surveying and Cultural Heritage
- Quantum-Dot Cellular Automata
- Anomaly Detection Techniques and Applications
- Speech and Audio Processing
- Bacterial Infections and Vaccines
- Machine Learning and Algorithms
Institute of Acoustics
2025
Chinese Academy of Sciences
2025
University of Chinese Academy of Sciences
2025
Baidu (China)
2023-2024
Harbin University of Science and Technology
2023
Yale University
2023
Central South University
2023
Xiangya Hospital Central South University
2023
University of Michigan
2022
Xidian University
2019-2020
Pedestrian detection is a crucial component for intelligent transport system and advanced driver assistance system. In recent years, pedestrian methods have achieved higher accuracy. However, the existing algorithms are insufficient small-scale that relatively far from cameras in practical applications. this paper, we propose novel deep sense network (termed SSN) detection. The proposed architecture could generate some proposal regions which more effective to detect pedestrians. Furthermore,...
Feature-aided tracking integrates supplementary features into traditional methods and improves the accuracy of data association that rely solely on kinematic measurements. However, previous applications feature-aided in multi-target passive sonar arrays directly utilized raw for likelihood calculations, causing performance degradation complex marine scenarios with low signal-to-noise ratio close-proximity trajectories. Inspired by successful application deep learning, this study proposes...
Meningitis is a severe and fatal neurological disease causes lots of burden. The purpose this study was to assess the global, regional, national burdens trends meningitis by age, sex, etiology.
Blockchain is multi-centralized, immutable and traceable, thus very suitable for distributed storage, privacy security management in IoTs. However, most existing researches focus on the integration of public blockchain In fact, problems such as slow consensus, low transmission throughput, completely open storage are intolerable IoT scenarios. Although consortium represented by Hyperledger Fabric has improved rate, its data relies PKI-based certificate mechanism, resulting inefficiency...
Vision Transformer and its variants have demonstrated great potential in various computer vision tasks. But conventional transformers often focus on global dependency at a coarse level, which suffer from learning challenge relationships fine-grained representation token level. In this paper, we introduce Multi-scale Attention Fusion into transformer (MAFormer), explores local aggregation feature extraction dual-stream framework for visual recognition. We develop simple but effective module...
Abstract The flatness is a key indicator to evaluate the riveting quality of countersunk rivets on skin aircraft or rockets. Aiming at deficiencies current measurement methods used by aerospace manufacturers, method rivet based visual 3D reconstruction proposed. Through zonal modulation structured light intensity projector, image interference problem caused highly reflective and rockets solved. A detection algorithm an arc feature designed realize accurate identification in image. surface...
A Colon Cleanliness Rating Algorithm (CCRA) based on colonoscopy image analysis is proposed in this paper, order to solve the problem that results of (or Bowel Preparation Quality) rating caused by manual inspection are inconsistent.Firstly, CCRA intercepts images from video. Secondly, each image's stool area segmented U-Net obtain 2-classification segmentation results. Finally, colon cleanliness obtained comparing average with standard proportion.After testing, pixel accuracy model 97.02 %,...
We present a novel masked image modeling (MIM) approach, context autoencoder (CAE), for self-supervised representation pretraining. pretrain an encoder by making predictions in the encoded space. The pretraining tasks include two tasks: prediction - predict representations patches, and patch reconstruction reconstruct patches. network is encoder-regressor-decoder architecture: takes visible patches as input; regressor predicts of which are expected to be aligned with computed from encoder,...
Change-Encryption is a distributed layer-by-layer symmetric encryption technology combined with blockchain and cryptography. The approach relies on the identity information of all nodes in blockchain, time, space, other attributes are used as parameters to enable repeated encryption. In this work, data sender receiver considered key, spatial values node dynamic increase uncertainty decryption, thereby solving security problem during communication. addition, self-identification principle...
Considering Grover's Search Algorithm (GSA) with the standard diffuser stage applied, we revisit $3$-qubit unique String Detection Problem (SDP) and extend algorithm to $4$-qubit SDP multiple winners. We then investigate unstructured search problems non-uniform distributions define Orthogonal Vector (OVP) under quantum settings. Although no numerically stable results is reached original GSA framework, provide intuition behind our implementation further observations on OVP. perform a special...
Objective A Colon Cleanliness Rating Algorithm(CCRA) based on colonoscopy image analysis is proposed in this paper, order to solve the problem that results of Cleanliness(or Bowel Preparation Quality) rating caused by manual inspection are inconsistent. Methods Firstly, CCRA intercepts images from video. Secondly, each image’s stool area segmented U-Net obtain 2-classification segmentation results. Finally, colon cleanliness obtained comparing average with standard proportion. Results...
With the development of computer vision, visual servo control has been widely used in various fields. This article analyzed stability image system application, and provided theoretical practical data support to engineering researchers when using technology. The Jacobian matrix based on point features was deduced detail, a simulation model an ideal environment built show convergence process. A 6-DOF manipulator for application experiments, several measures improve were described existing...
Computer scientists should be aware that their work may have an impact on society. In effort to integrate ethics and justice with technical content in CS education, we designed a novel CS1 activity. Our goals were: leave typical learning unchanged, include open-ended questions while meeting the constraints of large course, pose challenging students perceive as relevant computing practice personal life. this poster, discuss details activity designed, which includes coding portion where write...