Results 11 to 20 of about 1,461,530 (380)
FIERY: Future Instance Prediction in Bird’s-Eye View from Surround Monocular Cameras [PDF]
IEEE International Conference on Computer Vision, 2021 Driving requires interacting with road agents and predicting their future behaviour in order to navigate safely. We present FIERY: a probabilistic future prediction model in bird’s-eye view from monocular cameras.
Anthony Hu +7 more
semanticscholar +1 more source
Recurrent Vision Transformers for Object Detection with Event Cameras [PDF]
Computer Vision and Pattern Recognition, 2022 We present Recurrent Vision Transformers (RVTs), a novel backbone for object detection with event cameras. Event cameras provide visual information with submillisecond latency at a high-dynamic range and with strong robustness against motion blur.
Mathias Gehrig, D. Scaramuzza
semanticscholar +1 more source
E-RAFT: Dense Optical Flow from Event Cameras [PDF]
International Conference on 3D Vision, 2021 We propose to incorporate feature correlation and sequential processing into dense optical flow estimation from event cameras. Modern frame-based optical flow methods heavily rely on matching costs computed from feature correlation.
Mathias Gehrig +3 more
semanticscholar +1 more source
GLAMR: Global Occlusion-Aware Human Mesh Recovery with Dynamic Cameras [PDF]
Computer Vision and Pattern Recognition, 2021 We present an approach for 3D global human mesh recovery from monocular videos recorded with dynamic cameras. Our approach is robust to severe and long-term occlusions and tracks human bodies even when they go outside the camera's field of view.
Ye Yuan +4 more
semanticscholar +1 more source
Time-Ordered Recent Event (TORE) Volumes for Event Cameras [PDF]
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021 Event cameras are an exciting, new sensor modality enabling high-speed imaging with extremely low-latency and wide dynamic range. Unfortunately, most machine learning architectures are not designed to directly handle sparse data, like that generated from
R. W. Baldwin +4 more
semanticscholar +1 more source
ORB-SLAM2: An Open-Source SLAM System for Monocular, Stereo, and RGB-D Cameras [PDF]
IEEE Transactions on robotics, 2016 We present ORB-SLAM2, a complete simultaneous localization and mapping (SLAM) system for monocular, stereo and RGB-D cameras, including map reuse, loop closing, and relocalization capabilities. The system works in real time on standard central processing
Raul Mur-Artal, J. D. Tardós
semanticscholar +1 more source
Ecology and Evolution, 2021 Enrollment in courses taught remotely in higher education has been on the rise, with a recent surge in response to a global pandemic. While adapting this form of teaching, instructors familiar with traditional face‐to‐face methods are now met with a new ...
Frank R. Castelli, Mark A. Sarvary
semanticscholar +1 more source
Events-To-Video: Bringing Modern Computer Vision to Event Cameras [PDF]
Computer Vision and Pattern Recognition, 2019 Event cameras are novel sensors that report brightness changes in the form of asynchronous "events" instead of intensity frames. They have significant advantages over conventional cameras: high temporal resolution, high dynamic range, and no motion blur.
Henri Rebecq +3 more
semanticscholar +1 more source
Fooling Automated Surveillance Cameras: Adversarial Patches to Attack Person Detection [PDF]
2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2019 Adversarial attacks on machine learning models have seen increasing interest in the past years. By making only subtle changes to the input of a convolutional neural network, the output of the network can be swayed to output a completely different result.
Simen Thys, W. V. Ranst, T. Goedemé
semanticscholar +1 more source
EV-FlowNet: Self-Supervised Optical Flow Estimation for Event-based Cameras [PDF]
Robotics: Science and Systems, 2018 Event-based cameras have shown great promise in a variety of situations where frame based cameras suffer, such as high speed motions and high dynamic range scenes. However, developing algorithms for event measurements requires a new class of hand crafted
A. Z. Zhu +3 more
semanticscholar +1 more source