A New Perspective on Domain Adaptive Semantic Segmentation: T2S-DA Study Background and Significance Semantic segmentation plays a crucial role in computer vision, but its performance often relies on extensive labeled data. However, acquiring labeled data is costly, especially in complex scenarios. To address this, many studies turn to synthetic da...

One-shot Generative Domain Adaptation in 3D GANs In recent years, Generative Adversarial Networks (GANs) have achieved remarkable progress in the field of image generation. While traditional 2D generative models exhibit impressive performance across various tasks, extending this technology to 3D domains (3D-aware image generation) remains challengi...

Deep Learning Explainability Research: A Perturbation-Based Evaluation Method for Attribution Maps Background and Motivation With the remarkable success of deep learning models across various tasks, there is growing attention on the interpretability and transparency of these models. However, while these models excel in accuracy, their decision-maki...

Report on the Study of Deep Learning-Based Video Stabilization Methods and the SynthStab Synthetic Dataset Background Introduction Digital video stabilization technology, which removes unnecessary vibrations and camera motion artifacts through software, is a critical component in modern video processing, particularly for amateur video shooting. How...

Multi-View Transformation Network (MVTN): New Advances in 3D Understanding Research Background and Motivation Recent advancements in deep learning for 3D data in computer vision have achieved significant success, particularly in tasks like classification, segmentation, and retrieval. However, effectively utilizing 3D shape information remains a cha...

Research on Cross-Scale Co-Occurrence Local Binary Pattern (CS-COLBP) for Image Classification Image classification is a key area in computer vision, with feature extraction being its core research focus. The Local Binary Pattern (LBP), due to its efficiency and descriptive power, has been widely used in tasks such as texture classification and fac...

GAN Inversion and Image Editing New Method: Warping the Residuals for Image Editing with StyleGAN Background and Research Problem Generative Adversarial Networks (GANs) have made remarkable progress in the field of image generation, enabling the synthesis and editing of high-quality images. StyleGAN models, known for their semantically interpretabl...

Background and Significance Object re-identification (Re-ID) is an essential task in computer vision aimed at identifying specific objects across different times and scenes. Driven by deep learning, particularly convolutional neural networks (CNNs), this field has made significant strides. However, the emergence of vision transformers has opened ne...

Use of Transcranial Oscillatory Stimulation to Reduce Cybersickness in Virtual Reality Background and Motivation Virtual Reality (VR) technology is increasingly integrated into fields such as work, healthcare, and entertainment. However, approximately 95% of VR users experience symptoms known as Cybersickness (CS), characterized by nausea, dizzines...

Deep Learning-Driven Protein Design: Generating Functional Protein Sequences Using Conditional Diffusion Models Proteins are at the core of life sciences research and applications, offering countless possibilities due to their diversity and functional complexity. With advancements in deep learning technologies, protein design has reached a new pinn...