Yu-Lun Liu

TED-4DGS: Temporally Activated and Embedding-based Deformation for 4DGS Compression

Cheng‐Yuan Ho, Jui‐Chiu Chiang, Yu-Lun Liu, Wen-Hsiao Peng

Building on the success of 3D Gaussian Splatting (3DGS) in static 3D scene representation, its extension to dynamic scenes-commonly referred to as 4DGS or dynamic 3DGS- has attracted increasing attention. However, designing more compact, efficient deformation schemes together with rate-distortion-optimized compression strategies …

tw (code pays fourni par la source)

RL-AWB: Deep Reinforcement Learning for Auto White Balance Correction in Low-Light Night-time Scenes

Yuan-Kang Lee, Kuan-Lin Chen, Chia-Che Chang, Yu-Lun Liu

Nighttime color constancy still remains a challenging problem in computational photography due to low-light noise and complex illumination conditions. We present RL-AWB, a novel framework combining statistical methods with deep reinforcement learning for nighttime white balance. Our method begins with a statistical …

Skyfall-GS: Synthesizing Immersive 3D Urban Scenes from Satellite Imagery

Jie-Ying Lee, Yi Liu, Wei‐Cheng Chang, C.J. Wu et autres

tw, es, us (code pays fourni par la source)

RL-AWB: Deep Reinforcement Learning for Auto White Balance Correction in Low-Light Nighttime Scenes

Yuan-Kang Lee, Kuan-Lin Chen, Chia-Che Chang, Yu-Lun Liu

tw (code pays fourni par la source)

StealthAttack: Robust 3D Gaussian Splatting Poisoning via Density-Guided Illusions

Bo-Hsu Ke, Yu-Lun Liu, Wei-Chen Chiu

3D scene representation methods like Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) have significantly advanced novel view synthesis. As these methods become prevalent, addressing their vulnerabilities becomes critical. We analyze 3DGS robustness against image-level poisoning attacks and propose a novel …

tw (code pays fourni par la source)

LongSplat: Robust Unposed 3D Gaussian Splatting for Casual Long Videos

Chin-Yang Lin, Cheng Sun, Fu-En Yang, Min-Hung Chen et autres

LongSplat addresses critical challenges in novel view synthesis (NVS) from casually captured long videos characterized by irregular camera motion, unknown camera poses, and expansive scenes. Current methods often suffer from pose drift, inaccurate geometry initialization, and severe memory limitations. To address these …

tw, gb (code pays fourni par la source)

Every Camera Effect, Every Time, All at Once: 4D Gaussian Ray Tracing for Physics-based Camera Effect Data Generation

Yi Liu, Yu‐Hsiang Hsu, I-Sheng Fang, Yu-Lun Liu et autres

Common computer vision systems typically assume ideal pinhole cameras but fail when facing real-world camera effects such as fisheye distortion and rolling shutter, mainly due to the lack of learning from training data with camera effects. Existing data generation approaches suffer from …

SpectroMotion: Dynamic 3D Reconstruction of Specular Scenes

Cheng-De Fan, Chen-Wei Chang, Yi Liu, Jie-Ying Lee et autres

We present SpectroMotion, a novel approach that combines 3D Gaussian Splatting (3DGS) with physically-based rendering (PBR) and deformation fields to reconstruct dynamic specular scenes. Previous methods extending 3DGS to model dynamic scenes have struggled to represent specular surfaces accurately. Our method addresses …

tw (code pays fourni par la source)

Thin 3D: an Innovative Approach to Ultra-Thin Wafer-Level Active Device Transfer Technology With Optimized Material and Thermal Solution for 3D Ic

Ting‐Yu Chen, Shie-Ping Chang, Bo-Jheng Shih, Zih-Yang Chen et autres

To enable high-density integration, advanced 3D stacking technology implements backside wafer thinning, reducing wafer thickness to$10 \mu \mathrm{m}$. However, this approach imposes constraints on signal routing and via density in bonded structures. This work proposes a novel ultra-thin wafer-level active device transfer …

tw (code pays fourni par la source)

Fabrication Optimization and Reliability Study of Hyper RDL (HRDL) Interposer for Advanced Packaging and Heterogeneous Integration

Chun-Ta Li, Yu-Lun Liu, Tzu-Han Sun, Chien-Kang Hsiung et autres

Owing to warpage issues for multi-layer interposer, the device performance is limited by the I/O density of the interposer. The HRDL interposer, developed using a low-temperature hybrid bonding and layer transfer technique, significantly improves warpage issues and increases I/O density compared to …

tw (code pays fourni par la source)

CorrFill: Enhancing Faithfulness in Reference-based Inpainting with Correspondence Guidance in Diffusion Models

Kuan‐Hung Liu, Cheng-Kun Yang, Min-Hung Chen, Yu-Lun Liu et autres

In the task of reference-based image inpainting, an additional reference image is provided to restore a damaged target image to its original state. The advancement of diffusion models, particularly Stable Diffusion, allows for simple formulations in this task. However, existing diffusion-based methods …

Hyper RDL (HRDL) Interposer by Layer Transfer Technology for 3D IC and Advanced Packaging

Yu-Lun Liu, Chien-Kang Hsiung, Chun-Ta Li, Tzu-Han Sun et autres

Interposer technology is essential for electrical connections between dies and components. Among silicon, glass, and re-distribution layer (RDL) interposers, RDL offers improved signal integrity and cost reduction. This study expands on Hyper RDL (HRDL) interposers, using layer transfer and low-temperature hybrid bonding …

tw (code pays fourni par la source)