A5068772612- Music and Audio Processing
- Fire dynamics and safety research
- Flame retardant materials and properties
- Music Technology and Sound Studies
- Speech and Audio Processing
- Combustion and Detonation Processes
- Advanced Steganography and Watermarking Techniques
- Generative Adversarial Networks and Image Synthesis
- Thermal and Kinetic Analysis
- Video Analysis and Summarization
- Digital Media Forensic Detection
- Human Pose and Action Recognition
- Multimodal Machine Learning Applications
- Fire Detection and Safety Systems
- Chaos-based Image/Signal Encryption
- Advanced Sensor and Control Systems
- Simulation and Modeling Applications
- Advanced Algorithms and Applications
- Advanced MRI Techniques and Applications
- Topic Modeling
- Industrial Technology and Control Systems
- Advanced Neural Network Applications
- Image and Video Stabilization
- Advanced Neuroimaging Techniques and Applications
- Functional Brain Connectivity Studies
Tsinghua University
2022-2025
Communication University of China
2013-2024
Nanjing University of Science and Technology
2011-2024
Harbin University of Science and Technology
2024
Xinjiang University
2022-2023
North China University of Science and Technology
2022-2023
Dalian Medical University
2022
Second Affiliated Hospital of Dalian Medical University
2022
Shanxi Medical University
2022
Tianjin University
2021
Abstract The production of traditional Peking opera facial masks often relies on hand painting by experienced painters, which restricts the inheritance and development this intangible cultural heritage. Current research mainly focuses digital reconstruction storage existing masks, while high-quality mask generation technology is still in an infancy stage. In paper, different deep learning frameworks are improved for features generating new can effectively promote creative application masks....
Audio-visual question answering aims to answer questions regarding both audio and visual modalities in a given video, has drawn increasing research interest recent years. However, there have been no appropriate datasets for this challenging task on videos real-life scenarios so far. They are either designed with containing only clues without taking any information into account, or considering restrictions specific scenarios, such as panoramic about music performances. In paper, overcome the...
Multimodal sentiment analysis (MSA) is important for quickly and accurately understanding people's attitudes opinions about an event. However, existing methods suffer from the dominant contribution of text modality in dataset; this called dominance. In context, we emphasize that weakening role MSA tasks. To solve above two problems, perspective datasets, first propose Chinese multimodal opinion-level intensity (CMOSI) dataset. Three different versions dataset were constructed: manually...
This study proposes a new generation network based on transformers and guided by the music theory to produce high-quality work. In this study, decoding block of transformer is used learn internal information single-track music, cross-track are amongst tracks different musical instruments. A reward proposed, which optimizes global local loss objective functions while training discriminating so that can provide reliable adjustment method for network. The combining cross entropy guide generator...
Effective delivery and controlled release of metallo-prodrugs with sustained activation rapid response feed the needs precise medicine in metal chemotherapeutics. However, gold-based anticancer drugs often suffer from detoxification binding extracellular transfer by sulfur-containing peptides. To address this challenge, we integrate a thiol-activated prodrug strategy newly prepared hypercoordinated carbon-centered gold(I) clusters (HCGCs) their photosensitization character to augment...
Solvent screening is pivotal in the optimization of reaction conditions for metal‐catalyzed reactions. However, to date little known about role solvent effects driving situ structural evolution metal‐containing species towards polynuclear organometallic cluster intermediates and determining their reactivity. We herein isolate two distinct thirteen‐membered silver isomers, Ag13‐A Ag13‐M, acetone methanol, respectively. Significantly, we demonstrate interconversion these isomers facilitated by...
Solvent screening is pivotal in the optimization of reaction conditions for metal‐catalyzed reactions. However, to date little known about role solvent effects driving situ structural evolution metal‐containing species towards polynuclear organometallic cluster intermediates and determining their reactivity. We herein isolate two distinct thirteen‐membered silver isomers, Ag13‐A Ag13‐M, acetone methanol, respectively. Significantly, we demonstrate interconversion these isomers facilitated by...
Existing outdoor three-dimensional (3D) object detection algorithms mainly use a single type of sensor, for example, only using monocular camera or radar point cloud. However, sensors are affected by light and lose depth information. When scanning distant an occluded object, the data collected short-range cloud sensor very sparse, which affects algorithm. To address above challenges, we design deep learning network that can combine texture information two-dimensional (2D) geometric 3D...
This paper deals with the determination of selected fi re properties spruce wood.It describes general characteristics wood, microscopic and macroscopic features.Broad application this material requires assessment its regarding aspects, being a cause in forests or due to frequent occurrence wildland res Slovak territory, fl ammable used building industry, furniture etc.This analyses following properties: ash-ignition temperature, spontaneousignition mass burning rate, ignitability exposed...
The Internet of Things (IoT) is a burgeoning paradigm that changes our lives greatly. In many cases, IoT devices are located in remote areas, which can't be served by terrestrial access networks. As consequence, Low Earth Orbit (LEO) satellite communication system may play an important evolution 5G, and becomes paramount importance for those scenes relying on its disadvantages. Additionally, the first integration 5G also provides more technical support system. Based analysis special...
A modified total variation MRI image denoising method is proposed in this paper. First, the removes noise K-space compressed sensing reconstruction. Then, removed data used as a partial frequency observation model. The shows better results than RecPF method, LDP TVCMRI and FCSA sparse tested against Shepp-Logan phantom real MR images corrupted by of different intensity level, it gives Signal-to-Noise Ratio (SNR), relative error (ReErr), structural similarity (SSIM) RecPF, LDP, TVCMRI, FCSA.