蕭旭峰

FIAP: Fisher Information-based Anchor Pruning for Scaffold-GS Compression

機器人碩士學位學程

三維高斯潑濺、錨點式高斯潑濺、模型壓縮、錨點剪枝、費雪資訊、新視角合成、3D Gaussian Splatting、Scaffold-GS、Model Compression、Anchor Pruning、Fisher Information、Novel View Synthesis

A Paired Comparison of Feature X-Coordinate Midranges Before and After Projected-Y Windowing

機器人碩士學位學程

關鍵詞:徑向畸變、部分觀察、樣本中程值、X 跨度、座標投影、radial distortion、partial observation、sample midrange、X-span、coordinate projection

Calibration-Free Multi-Fisheye Image Automatic Stitching System Based on Segmented Spherical Unwarping

資訊科學與工程研究所

魚眼影像、全景拼接、免校正、分段式球面展開、幾何對齊、網格變形、Fisheye Images、Panoramic Stitching、Calibration-Free、Segmented Spherical Unwrapping、Geometric Alignment、Grid Deformation

Enhancing Long Video Representation: An Adaptive Framework Integrated with Patch-Wise Neural Video Representation

機器人碩士學位學程

隱式神經表示法、斯皮爾曼等級相關係數、視訊壓縮、魏克生符號檢定、參數資源分配、場景轉換偵測、Implicit Neural Representation、Wilcoxon-Signed Rank Test、Spearman Correlation Coefficient、Rate-Distortion Optimization、Video Compression、Scene Change Detection

Bifurcated Attention-Convolution Fusion for Enhanced Hyperprior Learned Image Compression

多媒體工程研究所

影像處理、學習式影像壓縮、深度學習、混合式架構、卷積神經網路、Image Processing、Image Compression、Deep Learning、Convolutional Neural Network、Transformer

An Efficient Privacy-Preserving Network Coding Scheme Based on Lightweight Homomorphic Encryption

資訊安全研究所

網路編碼、同態加密、基於環面的快速全同態加密方案、隱私保護、編碼、Network Coding、Homomorphic Encryption、TFHE、Privacy Preserving、Encoding

Pruned Temporal Difference Networks with Excitation for Action Recognition

資訊科學與工程研究所

動作識別、深度學習、激勵機制、模型剪枝、深度神經網路、Action recognition、deep learning、excitation、model pruning、deep neural network

A Dual-MLP Approach for Enhanced Neural Video Representation with Spatial-Temporal Modulation and Network Compression

機器人碩士學位學程

模型修剪、隱式神經表示、影像壓縮、神經視頻表示、模型壓縮、條件隱式函數、model compression、model pruning、implicit neural representation、neural video representation、video compression、instance-conditioned implicit functions

A Hybrid Convolutional and Transformer Network for Salient Object Detection

網路工程研究所

RGB-D 顯著物體偵測、Transformer、混和性架構、孿生網絡、超像素池化、ROI 影像壓縮、RGB-D salient object detection、transformers、Hybrid network、Siamese network、Superpixel-pooling、ROI image compression

Saliency-guided Recurrent Auto-Encoder for Video Compression

工學院機器人碩士學位學程

視訊壓縮、圖像壓縮、顯著物偵測、深度學習、機器學習、深度類神經網路、卷積類神經網路、自動編碼器、人眼視覺系統、圖像質量評估、video compression、image compression、salient object detection、deep learning、machine learning、deep neural networks、convolutional neural networks、autoencoders、human vision system、image quality assessment

Compression Framework of Surveillance Videos Based on Dynamic ROI

資訊學院資訊學程

關注區域、物件偵測、深度學習、視訊壓縮、超解析度、Region of interest、Object detection、Deep learning、Video compression、Super-resolution

GOAT: Grouped Object Attention for Trajectory prediction in urban traffic using HD map

資訊科學與工程研究所・共同指導：林文杰

自動駕駛、軌跡預測、注意力機制、城市交通、高精地圖、Autonomous Driving、Trajectory Prediction、Attention、Urban Traffic、High Definition Map

A Study of Semi-Automatic Data Annotation System for Machine Learning

資訊學院資訊學程・共同指導：陳耀宗

自動標記、機器學習、自動化標記系統、電腦視覺、標記偏差、Data Annotation System、Machine Learning、Auto Labeling、Computer Vision、Bias

RGBD salient object detection with parallel multi-contextual models and short connections

多媒體工程研究所

顯著物體偵測、物體偵測、深度學習、機器學習、深度類神經網路、卷積類神經網路、全卷積網路、salient object detection、object detection、deep learning、machine learning、deep neural network、convolutional neural network、fully convolutional network

Perceptual Quality and Statistical Analysis of Immersive Visual Contents

多媒體工程研究所

3D 影⽚主觀評量、沉浸式環境、基於深度資訊所設計雜訊視訊、基於視覺顯著 性所設計雜訊視訊、⽴體視覺、Subjective stereoscopic 3-D video quality assessment、Immersive system、Depth based distortion、Foveation based distortion、Stereoscopic vision.

The architecture of random linear fountain codes with hybrid coding in multiple Galois fields

多媒體工程研究所

通道編碼、有限域、伽羅瓦域、內容傳遞網路、噴泉碼、網路編碼、channel coding、finite field、Galois field、content distribution network、CDN、fountain codes、LT codes、random linear network coding

Hierarchical Superpixels-to-Supervoxels Segmentation Approach

網路工程研究所

超像素、超體素、圖像切割、影片切割、superpixel、supervoxel、image segmentation、video segmentation、lazy random walks

Inter-Frame Coding with Quadtree Template Matching for HEVC

資訊學院資訊學程

樣版匹配、四元樹分割、高效率視訊編碼、畫面間估測、template matching prediction、quadtree partitioning、high efficiency video coding、inter-frame prediction、low-rank matrix recovery or completion、overlapped block motion compensation

A Mechanism for QoS Improvement over Wi-Fi based on Streaming Traffic Behavior

網路工程研究所・共同指導：陳耀宗

服務品質、無線區域網路、封包排程、流量分類、Quality of service、WLAN、Packet scheduling、Traffic classification

Performance Evaluation of Live Video Publishing on Mobile Devices

網路工程研究所・共同指導：陳耀宗

RTMP、HTTP、即時串流、行動裝置、串流協定、Live Streaming、Mobile Device、Streaming Protocol

Perceptual distortion equalization and mode decision for video coding

多媒體工程研究所

視訊壓縮、感官品質、感官品質恆定編碼、Video Compression、Perceptual Quality、Constant Perceptual-Quality Coding