Filters








1,768 Hits in 8.2 sec

Weakly Supervised Object Detection with 2D and 3D Regression Neural Networks [article]

Florian Dubost, Hieab Adams, Pinar Yilmaz, Gerda Bortsova, Gijs van Tulder, M. Arfan Ikram, Wiro Niessen, Meike Vernooij, Marleen de Bruijne
2020 arXiv   pre-print
We propose a new weakly supervised detection method using neural networks, that computes attention maps revealing the locations of brain lesions.  ...  For comparison, we modify state-of-the-art methods to compute attention maps for weakly supervised object detection, by using a global regression objective instead of the more conventional classification  ...  This research was funded by The Netherlands Organisation for Health Research and Development (ZonMw) Project 104003005, with additional support of Netherlands Organisation for Scientific Research (NWO)  ... 
arXiv:1906.01891v4 fatcat:xzjplmpj6zamnpuyfk6q2porda

Weakly Supervised Object Detection with 2D and 3D Regression Neural Networks

Florian Dubost, Hieab Adams, Pinar Yilmaz, Gerda Bortsova, Gijs van Tulder, M.Arfan Ikram, Wiro Niessen, Meike W. Vernooij, Marleen de Bruijne
2020 Medical Image Analysis  
We propose a new weakly supervised detection method using neural networks, that computes attention maps revealing the locations of brain lesions.  ...  For comparison, we modify state-of-the-art methods to compute attention maps for weakly supervised object detection, by using a global regression objective instead of the more conventional classification  ...  Acknowledgements This research was funded by The Netherlands Organisation for Health Research and Development (ZonMw) Project 1040 030 05, with additional support of Netherlands Organisation for Scientific  ... 
doi:10.1016/j.media.2020.101767 pmid:32674042 fatcat:oiimvoyq55blpahle3jljynugu

Weakly Supervised 3D Object Detection from Point Clouds [article]

Zengyi Qin, Jinglu Wang, Yan Lu
2020 arXiv   pre-print
Weakly supervised learning is a promising approach to reducing the annotation requirement, but existing weakly supervised object detectors are mostly for 2D detection rather than 3D.  ...  The source code and pretrained models are publicly available at https://github.com/Zengyi-Qin/Weakly-Supervised-3D-Object-Detection.  ...  Existing studies on weakly supervised learning of object detection mainly focus on 2D detection [5, 22, 23] .  ... 
arXiv:2007.13970v1 fatcat:2n6irzxxh5cj7mjckign3knlsq

NeRF-Pose: A First-Reconstruct-Then-Regress Approach for Weakly-supervised 6D Object Pose Estimation [article]

Fu Li, Hao Yu, Ivan Shugurov, Benjamin Busam, Shaowu Yang, Slobodan Ilic
2022 arXiv   pre-print
Then, we train a pose regression network to predict pixel-wise 2D-3D correspondences between images and the reconstructed model. At inference, the approach only needs a single image as input.  ...  To avoid these problems, we present a weakly-supervised reconstruction-based pipeline, named NeRF-Pose, which needs only 2D object segmentation and known relative camera poses during training.  ...  We first reconstruct the object as a NeRF-based network trained with weak labels. Then, we train a pose regression network to regress the dense image pixel (2D)-object model (3D) correspondences.  ... 
arXiv:2203.04802v1 fatcat:7wvhmrdthvdexkzqu4uyd5hv34

MonoGRNet: A General Framework for Monocular 3D Object Detection [article]

Zengyi Qin, Jinglu Wang, Yan Lu
2021 arXiv   pre-print
MonoGRNet decomposes the monocular 3D object detection task into four sub-tasks including 2D object detection, instance-level depth estimation, projected 3D center estimation and local corner regression  ...  We propose MonoGRNet for the amodal 3D object detection from a monocular image via geometric reasoning in both the observed 2D projection and the unobserved depth dimension.  ...  Weakly supervised object detection Most existing studies focus on 2D object detection, while weakly supervised 3D detection has not been extensively explored.  ... 
arXiv:2104.08797v1 fatcat:x27vv6c3frekxonpsn4764lqyi

Neural Mesh Refiner for 6-DoF Pose Estimation [article]

Di Wu, Yihao Chen, Xianbiao Qi, Yongjian Yu, Weixuan Chen, Rong Xiao
2020 arXiv   pre-print
This paper bridges the gap between 2D mask generation and 3D location prediction via a differentiable neural mesh renderer.  ...  However, due to a lack of geometrical scene understanding from the directly regressed pose estimation, there are misalignments between the rendered mesh from the 3D object and the 2D instance segmentation  ...  Then we leverage the weakly supervised signal of 2D mask from the dual head prediction to refine the regressed 6-DoF pose estimation via a differentiable neural 3D mesh renderer.  ... 
arXiv:2003.07561v3 fatcat:3kst5u2mn5dm3npmduyolkls4e

Towards Single 2D Image-Level Self-Supervision for 3D Human Pose and Shape Estimation

Junuk Cha, Muhammad Saqlain, Changhwa Lee, Seongyeong Lee, Seungeun Lee, Donguk Kim, Won-Hee Park, Seungryul Baek
2021 Applied Sciences  
Our framework inputs single 2D images, estimates human 3D meshes in the intermediate layers, and is trained to solve four types of self-supervision tasks (i.e., three image manipulation tasks and one neural  ...  In this paper, we propose a self-supervised learning framework for 3D human pose and shape estimation that does not require other forms of supervision signals while using only single 2D images.  ...  [25] proposed a framework composed of a 2D joint detector based on CNN and inferred 3D poses from the detected 2D joints.  ... 
doi:10.3390/app11209724 fatcat:xtfyxi4bqvgt5jblvz6pwcmyd4

Weakly-Supervised Multi-Face 3D Reconstruction [article]

Jialiang Zhang, Lixiang Lin, Jianke Zhu, Steven C.H. Hoi
2021 arXiv   pre-print
The experimental results indicate that our proposed approach is very promising on face alignment tasks without fully-supervision and pre-processing like detection and crop.  ...  The de-facto pipeline for estimating the parametric face model from an image requires to firstly detect the facial regions with landmarks, and then crop each face to feed the deep learning-based regressor  ...  Then, we present the objective function of the proposed neural network. Finally, we give the detailed implementation on training and inference.  ... 
arXiv:2101.02000v1 fatcat:hstsoyvc75ew7pk5salpxpy3vq

Learning Temporal 3D Human Pose Estimation with Pseudo-Labels [article]

Arij Bouazizi and Ulrich Kressel and Vasileios Belagiannis
2021 arXiv   pre-print
A temporal convolutional neural network is trained with the generated 3D ground-truth and the geometric multi-view consistency loss, imposing geometrical constraints on the predicted 3D body skeleton.  ...  We present a simple, yet effective, approach for self-supervised 3D human pose estimation. Unlike the prior work, we explore the temporal information next to the multi-view self-supervision.  ...  [26] propose a method to regress the 3D keypoints by incorporating the temporal information next to the adversarial objective. A network distillation is employed for additional supervision.  ... 
arXiv:2110.07578v1 fatcat:e3syxlsvbzhi7g7hz5z7hu675i

Can 3D Pose be Learned from 2D Projections Alone? [article]

Dylan Drover, Rohith MV, Ching-Hang Chen, Amit Agrawal, Ambrish Tyagi,, Cong Phuoc Huynh
2018 arXiv   pre-print
Results on Human3.6M dataset demonstrates that our approach outperforms many previous supervised and weakly supervised approaches.  ...  3D pose estimation from a single image is a challenging task in computer vision. We present a weakly supervised approach to estimate 3D pose points, given only 2D pose landmarks.  ...  [28] combine a regression network which estimates 2D and 3D poses with, temporal smoothing and a parameterized, kinematic skeleton fitting method to produce stable 3D skeletons across time.  ... 
arXiv:1808.07182v1 fatcat:how62ljyffg5dpwz4cduubzare

Deep Learning Methods for 3D Human Pose Estimation under Different Supervision Paradigms: A Survey

Dejun Zhang, Yiqi Wu, Mingyue Guo, Yilin Chen
2021 Electronics  
Based on this literature survey, it can be concluded that each branch of 3D human pose estimation starts with fully-supervised methods, and there is still much room for multi-person pose estimation based  ...  on other supervision methods from both image and video.  ...  [57] address the problem by utilizing mixed 2D and 3D labels in a unified neural network consisting of a 2D module and a 3D module.  ... 
doi:10.3390/electronics10182267 fatcat:ajnizu776ncpto3jvyh3zye2si

Survey on depth and RGB image-based 3D hand shape and pose estimation

Lin Huang, Boshen Zhang, Zhilin Guo, Yang Xiao, Zhiguo Cao, Junsong Yuan
2021 Virtual Reality & Intelligent Hardware  
With the availability of large-scale annotated hand datasets and the rapid developments of deep neural networks (DNNs), numerous DNN-based data-driven methods have been proposed for accurate and rapid  ...  In this study, we present a comprehensive survey of state-of-the-art 3D hand shape and pose estimation approaches using RGB-D cameras.  ...  Discriminative approaches: Weakly-supervised and semi-supervised learning The increased number of real-world datasets with complete 3D annotations has been attributed to the fast growth of 3D hand shape  ... 
doi:10.1016/j.vrih.2021.05.002 fatcat:4tbhftt3ira6fporaqlscqhsse

Leveraging Pre-Trained 3D Object Detection Models For Fast Ground Truth Generation [article]

Jungwook Lee, Sean Walsh, Ali Harakeh, Steven L. Waslander
2018 arXiv   pre-print
This paper introduces a novel ground truth generation method that combines human supervision with pretrained neural networks to generate per-instance 3D point cloud segmentation, 3D bounding boxes, and  ...  We use the KITTI 3D object detection dataset to evaluate the efficiency and the quality of our annotation scheme.  ...  Fig.1 shows an example of an object from the KITTI object detection benchmark [1] with ground truth annotations in 2D and in 3D.  ... 
arXiv:1807.06072v1 fatcat:6gl6dzeeyrg63lyhrvbnc37vna

2020 Index IEEE Transactions on Pattern Analysis and Machine Intelligence Vol. 42

2021 IEEE Transactions on Pattern Analysis and Machine Intelligence  
., and Nishino, K., Recognizing Material Properties from Images; 1981-1995 Sebe, N., see Pilzer, A., 2380-2395 Seddik, M., see Tamaazousti, Y., 2212-2224 Shah, M., see Kalayeh, M.M., TPAMI June 2020  ...  Liu, R., +, TPAMI Dec. 2020 3027-3039 PCL: Proposal Cluster Learning for Weakly Supervised Object Detection.  ...  ., +, TPAMI June 2020 1515-1521 LCR-Net++: Multi-Person 2D and 3D Pose Detection in Natural Images.  ... 
doi:10.1109/tpami.2020.3036557 fatcat:3j6s2l53x5eqxnlsptsgbjeebe

Weakly Supervised 3D Human Pose and Shape Reconstruction with Normalizing Flows [article]

Andrei Zanfir, Eduard Gabriel Bazavan, Hongyi Xu, Bill Freeman, Rahul Sukthankar, Cristian Sminchisescu
2020 arXiv   pre-print
Monocular 3D human pose and shape estimation is challenging due to the many degrees of freedom of the human body and thedifficulty to acquire training data for large-scale supervised learning in complex  ...  In this paper we present practical semi-supervised and self-supervised models that support training and good generalization in real-world images and video.  ...  Unfortunately, predicting a 3d translation directly is difficult with neural networks.  ... 
arXiv:2003.10350v2 fatcat:c6gh2fxydve6hktgemewiutl4e
« Previous Showing results 1 — 15 out of 1,768 results