A5106716389- Forensic and Genetic Research
- Speech and Audio Processing
- Advanced Research in Systems and Signal Processing
- Handwritten Text Recognition Techniques
- Image Processing and 3D Reconstruction
- Geotechnical and Geomechanical Engineering
- Emotion and Mood Recognition
- Speech Recognition and Synthesis
- Music and Audio Processing
- Voice and Speech Disorders
- Vehicle License Plate Recognition
- Forensic Anthropology and Bioarchaeology Studies
Qilu University of Technology
2023
Shandong Academy of Sciences
2023
Automatic speaker verification (ASV) systems are vulnerable to synthetic speech attacks. Synthetic algorithms usually introduce artifacts in specific sub-bands or time segments. However, under unknown spoofing attacks, it is challenging choose the right domain for effective detection. In this paper, we propose a detection method based on graph attention networks with spectral and temporal information. First, high-level features of raw audio extracted using SENet channel enhance spatial...
The principles of operation the advanced resonant-reflectometric method for detecting GPS trackers and its advantages are described. It is shown that detection re-radiated signal depends on parameters such as power level duration probing pulses, frequencies pulses relative to resonance frequency tracker's input tract; quality factor tract distance from scanning antenna tracker. A block diagram experimental validation determination optimal proposed, a component base implementation selected....
How to identify bloodstains and obtain some potential evidence is of great significance for solving criminal cases. First, the spectral data different species bloodstain samples (human blood animal blood) were acquired by using a hyperspectral imager. Then, an extreme learning machine (ELM) algorithm was used build training models samples. Meanwhile, two traditional support vector random forest classification algorithms also compared with ELM algorithm. The prediction results showed that...
The most common trace of evidence at a crime scene is blood.
Facial expressions and speech are the most natural common ways for humans to express their emotions. Automatic audio-visual emotion recognition has attracted a lot of attention in recent years. However, previous methods cannot effectively exploit complementarity between modalities when performing feature fusion, resulting fused representations may contain redundant information from different modalities. This paper proposes an model based on multi-scale channel (MCA) global interactive fusion...