CVPR2022論文速遞（2022.6.7）！共14篇！
Updated on : ?7 Jun 2022
total number : 14
語義分割/Segmentation - 2 篇
Point-to-Voxel Knowledge Distillation for LiDAR Semantic Segmentation
標(biāo)題：激光雷達(dá)語義分割的點到素知識蒸餾
論文/Paper: http://arxiv.org/pdf/2206.02099
代碼/Code: https://github.com/cardwing/Codes-for-PVKD
Occlusion-Resistant Instance Segmentation of Piglets in Farrowing Pens Using Center Clustering Network
標(biāo)題：使用中心聚類網(wǎng)絡(luò)對豬的遮擋實例分割
論文/Paper: http://arxiv.org/pdf/2206.01942
代碼/Code: None
Transformers - - 2 篇
Scaling Vision Transformers to Gigapixel Images via Hierarchical Self-Supervised Learning
標(biāo)題：通過等級自監(jiān)督學(xué)習(xí)將視覺Transformers縮放到吉吉像素圖像
論文/Paper: http://arxiv.org/pdf/2206.02647
代碼/Code: None
Cross-modal Clinical Graph Transformer for Ophthalmic Report Generation
標(biāo)題：眼科報告生成的跨模式臨床圖Transformers
論文/Paper: http://arxiv.org/pdf/2206.01988
代碼/Code: None
姿態(tài)估計/Pose Estimation - 1 篇
Nerfels: Renderable Neural Codes for Improved Camera Pose Estimation
標(biāo)題：Nerfels：可渲染的神經(jīng)代碼，用于改進(jìn)相機(jī)姿勢估計
論文/Paper: http://arxiv.org/pdf/2206.01916
代碼/Code: None
人臉相關(guān) / Face - 2 篇
CORE: Consistent Representation Learning for Face Forgery Detection
標(biāo)題：CORE：面部偽造檢測的一致表示學(xué)習(xí)
論文/Paper: http://arxiv.org/pdf/2206.02749
代碼/Code: None
Evaluation-oriented Knowledge Distillation for Deep Face Recognition
標(biāo)題：以評估為導(dǎo)向的知識蒸餾以進(jìn)行深度識別
論文/Paper: http://arxiv.org/pdf/2206.02325
代碼/Code: None
其他/Other - 7 篇
Universal Photometric Stereo Network using Global Lighting Contexts
標(biāo)題：使用全局照明環(huán)境的通用光度立體網(wǎng)絡(luò)
論文/Paper: http://arxiv.org/pdf/2206.02452
代碼/Code: None
Invariant Grounding for Video Question Answering
標(biāo)題：視頻問答不變的基礎(chǔ)
論文/Paper: http://arxiv.org/pdf/2206.02349
代碼/Code: None
M2FNet: Multi-modal Fusion Network for Emotion Recognition in Conversation
標(biāo)題：M2FNet：在對話中識別情緒識別的多模式融合網(wǎng)絡(luò)
論文/Paper: http://arxiv.org/pdf/2206.02187
代碼/Code: None
MotionCNN: A Strong Baseline for Motion Prediction in Autonomous Driving
標(biāo)題：MotionCNN：自動駕駛中運動預(yù)測的強(qiáng)大基線
論文/Paper: http://arxiv.org/pdf/2206.02163
代碼/Code: None
Cannot See the Forest for the Trees: Aggregating Multiple Viewpoints to Better Classify Objects in Videos
標(biāo)題：看不到樹木的森林：匯總多個觀點以更好地對視頻中的對象進(jìn)行分類
論文/Paper: http://arxiv.org/pdf/2206.02116
代碼/Code: None
Learning sRGB-to-Raw-RGB De-rendering with Content-Aware Metadata
標(biāo)題：學(xué)習(xí)srgb-to-raw-rgb用內(nèi)容感知元數(shù)據(jù)
論文/Paper: http://arxiv.org/pdf/2206.01813
代碼/Code: https://github.com/SamsungLabs/content-aware-metadata)
RIDDLE: Lidar Data Compression with Range Image Deep Delta Encoding
標(biāo)題：RIDDLE：帶有范圍圖像的LIDAR數(shù)據(jù)壓縮深層增量編碼
論文/Paper: http://arxiv.org/pdf/2206.01738
代碼/Code: None