A5044278717- Visual Attention and Saliency Detection
- Surface Treatment and Residual Stress
- Aesthetic Perception and Analysis
- Multimodal Machine Learning Applications
- Fatigue and fracture mechanics
- Language, Metaphor, and Cognition
- Erosion and Abrasive Machining
- Image and Video Quality Assessment
- Image Retrieval and Classification Techniques
- High-Velocity Impact and Material Behavior
- Image Enhancement Techniques
- Subtitles and Audiovisual Media
- Metal Forming Simulation Techniques
- Advanced Measurement and Detection Methods
- Advanced Vision and Imaging
- Metallurgy and Material Forming
- 3D Surveying and Cultural Heritage
- Advanced Image and Video Retrieval Techniques
- Laser and Thermal Forming Techniques
Xidian University
2023-2024
South China University of Technology
2012-2020
With collective endeavors, multimodal large language models (MLLMs) are undergoing a flourishing development. However, their performances on image aesthetics perception remain indeterminate, which is highly desired in real-world applications. An obvious obstacle lies the absence of specific benchmark to evaluate effectiveness MLLMs aesthetic perception. This blind groping may impede further development more advanced with capacity. To address this dilemma, we propose AesBench, an expert...
Image aesthetics assessment (IAA) aims at predicting the aesthetic quality of images. Recently, large pre-trained vision-language models, like CLIP, have shown impressive performances on various visual tasks. When it comes to IAA, a straightforward way is finetune CLIP image encoder using However, this can only achieve limited success without considering uniqueness multimodal data in domain. People usually assess according fine-grained attributes, e.g., color, light and composition. how...
The highly abstract nature of image aesthetics perception (IAP) poses significant challenge for current multimodal large language models (MLLMs). lack human-annotated multi-modality aesthetic data further exacerbates this dilemma, resulting in MLLMs falling short capabilities. To address the above challenge, we first introduce a comprehensively annotated Aesthetic Multi-Modality Instruction Tuning (AesMMIT) dataset, which serves as footstone building foundation models. Specifically, to align...
Shot peening is one of the most effective surface strengthening treatment method. Residual compressive stresses occur on workpiece after shot peening, and results in increase fatigue life workpiece. Finite element analysis (FEA) method was usually used to simulate distributions residual during due high cost long time needed by experimental research, random finite model normally adopted process based characteristics motion shots position indentations. Coverage an important parameter measure...
Abstract The aircraft corroded components can continue to be used after removing the portion by grinding and strengthening ground area shot peening. finite element analysis model of specimen with complex shape surface was established based on ABAQUS software. residual stress distributions under three conditions without peening, peening surfaces were analysed. On this basis, cyclic alternating load introduced into model, fatigue carried out MSC. Fatigue result shows that position maximum...