A5100415324- Biometric Identification and Security
- Face recognition and analysis
- Forensic Fingerprint Detection Methods
- Face and Expression Recognition
- Advanced Image and Video Retrieval Techniques
- Speech and Audio Processing
- Video Surveillance and Tracking Methods
- Digital Media Forensic Detection
- Forensic and Genetic Research
- Advanced Vision and Imaging
- Generative Adversarial Networks and Image Synthesis
- Remote Sensing and LiDAR Applications
- Gait Recognition and Analysis
- Image Retrieval and Classification Techniques
- Remote Sensing and Land Use
- Anomaly Detection Techniques and Applications
- Advanced Steganography and Watermarking Techniques
- Advanced Malware Detection Techniques
- Optical Coherence Tomography Applications
- Speech Recognition and Synthesis
- Wood and Agarwood Research
- Chaos-based Image/Signal Encryption
- Advanced Algorithms and Applications
- Network Security and Intrusion Detection
- Adversarial Robustness in Machine Learning
Nanjing University of Science and Technology
1994-2025
Sichuan University
2013-2025
Yunnan Academy of Forestry
2023-2025
Stomatology Hospital
2025
State Key Laboratory of Oral Diseases
2025
Shenzhen University
2015-2024
China National Petroleum Corporation (China)
2024
Huawei Technologies (Germany)
2024
Institute of Microelectronics
2006-2024
Xi'an Polytechnic University
2024
Detecting manipulated facial images and videos is an increasingly important topic in digital media forensics. As advanced face synthesis manipulation methods are made available, new types of fake representations being created which have raised significant concerns for their use social media. Hence, it crucial to detect localize regions. Instead simply using multi-task learning simultaneously predict the mask (regions), we propose utilize attention mechanism process improve feature maps...
Spatio-temporal prediction is a key type of tasks in urban computing, e.g., traffic flow and air quality. Adequate data usually prerequisite, especially when deep learning adopted. However, the development levels different cities are unbalanced, still many suffer from scarcity. To address problem, we propose novel cross-city transfer method for spatio-temporal tasks, called RegionTrans. RegionTrans aims to effectively knowledge data-rich source city data-scarce target city. More...
The identity of the earliest inhabitants Xinjiang, in heart Inner Asia, and languages that they spoke have long been debated remain contentious1. Here we present genomic data from 5 individuals dating to around 3000-2800 BC Dzungarian Basin 13 2100-1700 BC Tarim Basin, representing yet discovered human remains North South respectively. We find Early Bronze Age exhibit a predominantly Afanasievo ancestry with an additional local contribution, Early-Middle contain only ancestry. site Xiaohe...
This paper presents a stylized novel view synthesis method. Applying state-of-the-art stylization methods to views frame by often causes jittering artifacts due the lack of cross-view consistency. Therefore, this investigates 3D scene that provides strong inductive bias for consistent synthesis. Specifically, we adopt emerging neural radiance fields (NeRF) as our choice representation their capability render high-quality variety scenes. However, rendering from NeRF requires large number...
To address the COVID-19 pandemic, governments worldwide implemented mandatory restrictions. As an unintended consequence of these responses, significant air pollution reductions have been recorded across world. We provide cross-national evidence on causal impact pandemic-induced lockdowns quality. Using daily data between 1 January and 31 December 2020, covering 596 major cities in 77 countries, we analyzed with a generalized difference-in-differences approach. The results show that lockdown...
High-resolution automated fingerprint recognition systems (AFRSs) offer higher security because they are able to make use of level-3 features, such as pores, that not available in lower resolution ( <; 500-dpi) images. One the main parameters affecting quality a digital image and issues cost, interoperability, performance an AFRS is choice resolution. In this paper, we identify optimal for using two most representative features: minutiae pores. We first designed multiresolution acquisition...
Abstract Commercially available extra virgin olive oils are often adulterated with some other cheaper edible similar chemical compositions. A set of oil samples soybean oil, corn and sunflower seed were characterized by Raman spectra in the region 1000–1800 cm −1 . Based on intensity vibrational bands normalized band at 1441 (CH 2 ), external standard method (ESM) was employed for quantitative analysis, which compared results achieved support vector machine (SVM) methods. By plotting...
In this paper, we proposed a effective but extensible residual one-dimensional convolution neural network as base network, based on the four subnets to explore features of skeleton sequences from each aspect. Given sequences, spatial information are encoded into joints coordinate in frame and temporal present by multiple frames. Limited sequence representations, two-dimensional cannot be used directly, chose layer basic layer. Each sub could extract discriminative different aspects. Our...
Supervised learning aims to train a classifier under the assumption that training and test data are from same distribution. To ease above assumption, researchers have studied more realistic setting: out-of-distribution (OOD) detection, where may come classes unknown during (i.e., OOD data). Due unavailability diversity of data, good generalization ability is crucial for effective detection algorithms. study in this paper, we investigate probably approximately correct (PAC) theory which...
Due to the diversity of attack materials, fingerprint recognition systems (AFRSs) are vulnerable malicious attacks. It is thus important propose effective presentation detection (PAD) methods for safety and reliability AFRSs. However, current PAD often exhibit poor robustness under new types settings. This paper proposes a novel channel-wise feature denoising (CFD-PAD) method by handling redundant noise information ignored in previous studies. The proposed learns features images weighing...
Abstract Lip-to-Speech (LTS) generation is an emerging technology that highly visible, widely supported, and rapidly evolving. LTS has a wide range of promising applications, including assisting speech impairment improving interaction in virtual assistants robots. However, the technique faces following challenges: (1) Chinese lip-to-speech poorly recognized. (2) The variation lip-speaking aligned with lip movements. Addressing these challenges will contribute to advancing technology,...
Touchless fingerprint capture devices have the advantage over traditional touch-based approaches of being hygienic and preventing distortions resulting from contact fingers. Single-view acquisition systems bring in problems, such as scene difference a limited effective area. This paper thus presents touchless multiview system that acquires three different views images at same time. device is designed by optimizing parameters regarding captured image quality size. A mosaicking method proposed...