V
主页
20220907【自动驾驶感知】李鸿升:MPPNet: Multi-Frame Feature Intertwining with Proxy Points……
发布人
报告嘉宾:李鸿升 (香港中文大学) 报告时间:2022年09月07日 (星期三)晚上20:00 (北京时间) 报告题目:MPPNet: Multi-Frame Feature Intertwining with Proxy Points for 3D Temporal Object Detection 报告人简介: 李鸿升,现任香港中文大学多媒体实验室副教授和西安电子科技大学“华山学者”讲座教授,研究方向为计算机视觉、深度学习以及医学图像处理。他在计算机视觉和医学图像处理顶级会议 (CVPR/ ICCV/ ECCV/ NeurlPS/ MICCAI)已经发表论文100余篇,Google Scholar引用超过21,000次。他作为团队负责人参加ImageNet 2016大赛,获得了视频物体检测第一名。他获得了2020年IEEE电路与系统协会杰出青年作者奖,2022年AI 2000计算机视觉全球最具影响力学者提名奖,2021年CUHK青年学者研究成就奖等奖项。他担任过NeurIPS 2021、2022领域主席和Neurocomputing的副编辑。 个人主页: https://www.ee.cuhk.edu.hk/~hsli/ 报告摘要: Accurate and reliable 3D detection is vital for many applications including autonomous driving vehicles and service robots. We present a flexible and high-performance 3D detection framework, named MPPNet, for 3D temporal object detection with point cloud sequences. We propose a novel three-hierarchy framework with proxy points for multi-frame feature encoding and interactions to achieve better detection. The three hierarchies conduct per-frame feature encoding, short-clip feature fusion, and whole-sequence feature aggregation, respectively. To enable processing long-sequence point clouds with reasonable computational resources, intra-group feature mixing and inter-group feature attention are proposed to form the second and third feature encoding hierarchies, which are recurrently applied for aggregating multi-frame trajectory features. The proxy points not only act as consistent object representations for each frame, but also serve as the courier to facilitate feature interaction between frames. The experiments on large Waymo Open dataset show that our approach outperforms state-of-the-art methods with large margins when applied to both short (e.g., 4-frame)and long (e.g., 16-frame)point cloud sequences. Specifically, MPPNet achieves 74.21%, 74.62% and 73.31% for vehicle, pedestrian and cyclist classes on the LEVEL 2 mAPH metric with 16-frame input.
打开封面
下载高清视频
观看高清视频
视频下载器
【VALSE2024】0505 吴小俊《特邀报告:多模态视觉融合方法:是否存在性能极限?》
20240828【医学多模态分析与研究:从传统模型到大模型的演变】王淑君:Multi-modal Representation Learning for Me…
【VALSE2024】0505 林倞《APR:面向具身智能的多模态感知与交互》
20240717【面向事件相机的物体检测与跟踪】李家宁:Object Detection with Neuromorphic Cameras
20211222【自动驾驶中的机器视觉与学习问题】赵行:环视自动驾驶感知
20211201【薪火相传经验谈——学长学姐教你如何做科研】闫杰熹:沟通,让科研变得美好
20220907【自动驾驶感知】Panel
20220309【让机器看懂视频:视频分割与目标追踪】杨宗鑫:视频理解中的多目标联合分割
20230607【开放世界的感知:探索可迁移与可持续学习之路】巩东:Continual Learning and Memory Augmentation……
【VALSE论文速览-94期】Learning Domain Adaptive Object Detection with Probabilistic……
【VALSE2024】0505 卢志武《APR:视频生成》
20220615【AI for Science之物理信息驱动的深度学习】陆路:Learning operators using deep neural……
【VALSE2024】0506《Workshop :具身智能的视觉与学习》
20220615【AI for Science之物理信息驱动的深度学习】王建勋:Leveraging physics-induced bias in……
20211201【薪火相传经验谈——学长学姐教你如何做科研】游凯超:我的科研经历与科研习惯
20220316【基于领域知识的机器学习在医学影像分析中的应用】崔智铭:Shape-aware tooth segmentation in digital……
20220413【脑启发视觉】Panel
20220921【“眼见未必为实”:人脸安全研究进展】邓伟洪:安全可信的人脸识别
【VALSE2024】0505 杨耀东《APR:从偏好对齐到价值对齐与超对齐》
20240605【Prompt Learning in Vision】刘东方:Prompt Tuning as Sustainable Fast Learner
20240814【多模态医学图像处理及医学大模型的发展近况】陈浩:多模态计算病理基准模型:挑战和未来
20230531【大模型时代下的三维视觉:路在何方?】杨波:3D Semantic and Instance Segmentation without 3D……
20210120【无人驾驶场景3D物体检测与跟踪】王哲 Multi-Modality Cut and Paste for 3D Object Detection
20220629【比物连类:对比表示学习】宫明明:CRIS: CLIP-Driven Referring Image Segmentation
20220831【就正有道:物理机理驱动的图像恢复与增强】Panel
20210421【无师自通:自监督学习】田渊栋:Understanding representation learning without negative……
20210120【无人驾驶场景3D物体检测与跟踪】 周星壹 Center-based 3D Object Detection and Tracking
【SAM2+YOLOV8】Meta首个能在图片和视频中实时分割对象的统一开源AI模型!博士精讲论文及源码复现!
20210625;短教程:《Transformers》;特邀讲师:邱锡鹏教授
【VALSE论文速览-86期】OFA: Unifying Architectures, Tasks, and Modalities Through a……
20211124【标记高效的视觉学习】Panel
20220629【比物连类:对比表示学习】Panel
【VALSE2024】0505 王兴刚《APR:面向大模型的新型高效率网络架构》
20240821【AI for Neuroscience, what and how?】李远宁:Enhancing neural encoding model…
整整200集,比刷剧还爽!深度学习必看圣经!李沐大神《动手学深度学习》最新版全套视频教程分享(人工智能/深度学习/计算机视觉)
【VALSE2024】0505 俞扬《APR:世界模型与具身决策》
20210818【心中的象牙塔:怎样才能拿到理想的教职offer?】陈键飞:学术道路上的一些感悟
20221123【深度网络交互学习:合作, 对抗与融合】徐行:Cross-modal Correlation Learning via Knowledge……
20240828【医学多模态分析与研究:从传统模型到大模型的演变】Panel
【VALSE2024】0505 杨易《特邀报告:混合模型驱动的内容生成与具身智能》