A5101571422- Human Pose and Action Recognition
- Hand Gesture Recognition Systems
- Anomaly Detection Techniques and Applications
- Context-Aware Activity Recognition Systems
- Ergonomics and Musculoskeletal Disorders
- Privacy-Preserving Technologies in Data
- Speech and Audio Processing
- Video Surveillance and Tracking Methods
- Time Series Analysis and Forecasting
- Internet Traffic Analysis and Secure E-voting
- Speech Recognition and Synthesis
- Advanced Steganography and Watermarking Techniques
- Advanced Chemical Sensor Technologies
- Music and Audio Processing
- Gait Recognition and Analysis
Fraunhofer Institute for Digital Media Technology
2024
Northeastern University
2021
Nanchang University
2017-2020
Automatic human fall detection is one important research topic in caring for vulnerable people, such as elders at home and patients medical places. Over the past decade, numerous methods aiming solving problem were proposed. However, existing only focus on detecting themselves cannot work effectively complicated environments, especially falls furniture. To alleviate this problem, a new method furniture using scene analysis based deep learning activity characteristics presented paper. The...
The existing deep learning methods for human fall detection have difficulties to distinguish falls from similar daily activities such as lying down because of not using the 3D network. Meanwhile, they are suitable mobile devices heavyweight and consume a large number memories. In order alleviate these problems, two-stream approach with MobileVGG is proposed in this paper. One stream based on motion characteristics body falls, while other an improved lightweight VGG network, named MobileVGG,...
Behavior analysis through posture recognition is an essential research in robotic systems. Sitting with unhealthy sitting for a long time seriously harms human health and may even lead to lumbar disease, cervical disease myopia. Automatic vision-based detection of posture, as example systems, has become hot topic. However, the existing methods only focus on extracting features themselves lack understanding relevancies among objects scene, henceforth fail recognize some types postures...
Research on speaker recognition is extending to address the vulnerability in wild conditions, among which genre mismatch perhaps most challenging, for instance, enrollment with reading speech while testing conversational or singing audio. This leads complex and composite inter-session variations, both intrinsic (i.e., speaking style, physiological status) extrinsic recording device, background noise). Unfortunately, few existing multi-genre corpora are not only limited size but also recorded...
In this paper, we introduce FLCrypt <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> For evaluation access to FLCrypt, please contact the main author or reach us via https://www.idmt.fraunhofer.de/en/contact.html., a library designed enhance Federated Learning with additional privacy guarantees by applying Fully Homomorphic Encryption model aggregation stage, thereby preventing aggregator from accessing unencrypted parameters of training...