A5070637775- Anomaly Detection Techniques and Applications
- Context-Aware Activity Recognition Systems
- Rough Sets and Fuzzy Logic
- Machine Learning and Data Classification
- IoT and Edge/Fog Computing
- Non-Invasive Vital Sign Monitoring
- Multi-Criteria Decision Making
- Imbalanced Data Classification Techniques
- Human Pose and Action Recognition
- Cognitive Science and Mapping
Southeast University
2023
Human activity recognition (HAR) based on sensor information has become a hot topic of research due to its wide range applications in health care, fitness, and smart homes. However, the classification activities with similar signals such as standing sitting is usually more challenging for design efficient algorithms. Considering characteristic human different granularity, which can provide complementary knowledge individual granularity recognition, we propose novel approach that combines...
The belief functions (BFs) introduced by Shafer in the mid of 1970s are widely applied information fusion to model epistemic uncertainty and reason about uncertainty. Their success applications is however limited because their high-computational complexity process, especially when number focal elements large. To reduce reasoning with BFs, we can envisage as a first method involved process convert original basic assignments (BBAs) into simpler ones, or second use simple rule combination...
Human Activity Recognition (HAR) based on wear-able device has become a hot topic of research due to its wide range applications in health-care, fitness and smart homes. However, the classification some activities with similar sensor readings, such as standing sitting, is usually more challenging for design efficient activity recognition algorithms. Considering inconsistent performance different classifiers, which can provide information complementary individual classifier, we propose novel...
Dezert-Smarandache Theory (DSmT) can effectively model sensor information and offer combination rules for multi-source fusion. However, fusion results obtained using DSmT are often unsatisfactory when a particular source is unreliable. To improve the reliability of results, new with multi-criteria evaluation (FMCE) proposed. Firstly, we evaluate from two aspects: similarity imprecision. Then, The used to calculate appropriate weighting coefficient each source. Next, belief values...