Filters








145 Hits in 4.0 sec

Learning Transferrable Knowledge for Semantic Segmentation with Deep Convolutional Neural Network [article]

Seunghoon Hong, Junhyuk Oh, Bohyung Han, Honglak Lee
2015 arXiv   pre-print
We propose a novel weakly-supervised semantic segmentation algorithm based on Deep Convolutional Neural Network (DCNN).  ...  To make the segmentation knowledge transferrable across categories, we design a decoupled encoder-decoder architecture with attention model.  ...  Recent breakthrough in semantic segmentation has been mainly accelerated by the approaches based on Convolutional Neural Networks (CNNs) [21, 4, 11, 10, 25] .  ... 
arXiv:1512.07928v1 fatcat:p6kdgj7gbvdrrgxzrdhss275k4

Learning Transferrable Knowledge for Semantic Segmentation with Deep Convolutional Neural Network

Seunghoon Hong, Junhyuk Oh, Honglak Lee, Bohyung Han
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
We propose a novel weakly-supervised semantic segmentation algorithm based on Deep Convolutional Neural Network (DCNN).  ...  To make segmentation knowledge transferrable across categories, we design a decoupled encoder-decoder architecture with attention model.  ...  Recent breakthrough in semantic segmentation has been mainly accelerated by the approaches based on Convolutional Neural Networks (CNNs) [4, 21, 11, 10, 25] .  ... 
doi:10.1109/cvpr.2016.349 dblp:conf/cvpr/HongOLH16 fatcat:obemojsyvvfflodeaxd3ptfyse

Transfer Learning from Audio-Visual Grounding to Speech Recognition

Wei-Ning Hsu, David Harwath, James Glass
2019 Interspeech 2019  
Transfer learning aims to reduce the amount of data required to excel at a new task by re-using the knowledge acquired from learning other related tasks.  ...  As semantics of speech are largely determined by its lexical content, grounding models learn to preserve phonetic information while disregarding uncorrelated factors, such as speaker and channel.  ...  Deep audiovisual embedding network (DAVEnet) [21] is a two-branched convolutional neural network model for this task, which learns to encode images and spoken captions into a shared embedding space that  ... 
doi:10.21437/interspeech.2019-1227 dblp:conf/interspeech/HsuHG19 fatcat:vtf6iei6xfh5rela5wa77r6brq

CATCH: Context-based Meta Reinforcement Learning for Transferrable Architecture Search [article]

Xin Chen, Yawen Duan, Zewei Chen, Hang Xu, Zihao Chen, Xiaodan Liang, Tong Zhang, Zhenguo Li
2020 arXiv   pre-print
These challenges preclude their applicability, and motivate our proposal of CATCH, a novel Context-bAsed meTa reinforcement learning (RL) algorithm for transferrable arChitecture searcH.  ...  The contexts also assist a network evaluator in filtering inferior candidates and speed up learning.  ...  Introduction The emergence of many high-performance neural networks has been one of the pivotal forces pushing forward the progress of deep learning research and production.  ... 
arXiv:2007.09380v3 fatcat:3mqj2rjlgvaprpxr6knxlx5x44

Towards Single Stage Weakly Supervised Semantic Segmentation [article]

Peri Akiva, Kristin Dana
2021 arXiv   pre-print
The costly process of obtaining semantic segmentation labels has driven research towards weakly supervised semantic segmentation (WSSS) methods, using only image-level, point, or box labels.  ...  The multi-stage approach is computationally expensive, and dependency on image-level labels for CAMs generation lacks generalizability to more complex scenes.  ...  Learning transferrable knowledge for semantic seg- Hsuan Yang. Weakly supervised object localization with mentation with deep convolutional neural network.  ... 
arXiv:2106.10309v2 fatcat:l3oafc7rz5frbiynru2vn6ogfa

Snapshot Distillation: Teacher-Student Optimization in One Generation

Chenglin Yang, Lingxi Xie, Chi Su, Alan L. Yuille
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Optimizing a deep neural network is a fundamental task in computer vision, yet direct training methods often suffer from over-fitting.  ...  We also verify that models pre-trained with SD transfers well to object detection and semantic segmentation in the PascalVOC dataset.  ...  Introduction A large portion of recent advances in computer vision have been built upon deep learning, in particular training very deep neural networks.  ... 
doi:10.1109/cvpr.2019.00297 dblp:conf/cvpr/YangXSY19 fatcat:sbd6ifmqhbe4hfoeekllcrnprm

Snapshot Distillation: Teacher-Student Optimization in One Generation [article]

Chenglin Yang, Lingxi Xie, Chi Su, Alan L. Yuille
2018 arXiv   pre-print
Optimizing a deep neural network is a fundamental task in computer vision, yet direct training methods often suffer from over-fitting.  ...  We also verify that models pre-trained with SD transfers well to object detection and semantic segmentation in the PascalVOC dataset.  ...  Introduction A large portion of recent advances in computer vision have been built upon deep learning, in particular training very deep neural networks.  ... 
arXiv:1812.00123v1 fatcat:o7fpnqeulffshgsbgvpjvydcfi

Geometry Guided Convolutional Neural Networks for Self-Supervised Video Representation Learning

Chuang Gan, Boqing Gong, Kun Liu, Hao Su, Leonidas J. Guibas
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
Although the geometry and highlevel semantics are seemingly distant topics, surprisingly, we find that the convolutional neural networks pre-trained by the geometry cues can be effectively adapted to semantic  ...  noisy and indirect training signals for learning the video representations.  ...  To recognize actions and events happening in videos, recent approaches that employ deep convolutional neural networks (CNNs) [12, 17, 31, 34, 35] , recurrent networks [15, 33, 4] , and attention networks  ... 
doi:10.1109/cvpr.2018.00586 dblp:conf/cvpr/GanGLSG18 fatcat:fris6ki3zbasdd7t36jomx24hy

Knowledge Distillation in Generations: More Tolerant Teachers Educate Better Students [article]

Chenglin Yang, Lingxi Xie, Siyuan Qiao, Alan Yuille
2018 arXiv   pre-print
We focus on the problem of training a deep neural network in generations.  ...  Consequently, the teacher provides a milder supervision signal (a less peaked distribution), and makes it possible for the student to learn from inter-class similarity and potentially lower the risk of  ...  This research votes for the viewpoint that network optimization is far from perfect at the current status.  ... 
arXiv:1805.05551v2 fatcat:3e2yziuwurbhtfdfn25jzzhpjq

Concept Representation Learning with Contrastive Self-Supervised Learning [article]

Daniel T. Chang
2022 arXiv   pre-print
Concept-oriented deep learning (CODL) is a general approach to meet the future challenges for deep learning: (1) learning with little or no external supervision, (2) coping with test examples that come  ...  from a different distribution than the training examples, and (3) integrating deep learning with symbolic AI.  ...  Acknowledgement: Thanks to my wife Hedy (郑期芳) for her support.  ... 
arXiv:2112.05677v2 fatcat:z2yfbtrsvzd4ractfjrechopqa

Training Constrained Deconvolutional Networks for Road Scene Semantic Segmentation [article]

German Ros, Simon Stent, Pablo F. Alcantarilla, Tomoki Watanabe
2016 arXiv   pre-print
In this work we investigate the problem of road scene semantic segmentation using Deconvolutional Networks (DNs).  ...  of knowledge from S-Net to the memory-efficient target network (T-Net).  ...  With the arrival of deep convolutional neural networks (CNNs), hand-crafted features were substituted by learned CNN representations, which worked at the level of image patches [23, 24] .  ... 
arXiv:1604.01545v1 fatcat:flm7e6sd35av7c45ycvr7zx5nu

Attacking Point Cloud Segmentation with Color-only Perturbation [article]

Jiacen Xu, Zhe Zhou, Boyuan Feng, Yufei Ding, Zhou Li
2021 arXiv   pre-print
Recent research efforts on 3D point-cloud semantic segmentation have achieved outstanding performance by adopting deep CNN (convolutional neural networks) and GCN (graph convolutional networks).  ...  Given that semantic segmentation has been applied in many safety-critical applications (e.g., autonomous driving, geological sensing), it is important to fill this knowledge gap, in particular, how these  ...  Kpconv: Flexible and Minimal adversarial examples for deep learning on 3d point deformable convolution for point clouds.  ... 
arXiv:2112.05871v2 fatcat:rru4yw6drjbrvpboxaruqo7asa

Deep Learning for Political Science [article]

Kakia Chatsiou, Slava Jankin Mikhaylov
2020 arXiv   pre-print
The discussion of deep neural networks is illustrated with the NLP tasks that are relevant to political science.  ...  The latest advances in deep learning methods for NLP are also reviewed, together with their potential for improving information extraction and pattern recognition from political science texts.  ...  transfer learning, we are also further understanding what we are learning with the deep neural networks.  ... 
arXiv:2005.06540v1 fatcat:kz2cbxjrmrfhdlfss5gqziefoq

Artificial Intelligence in Quantitative Ultrasound Imaging: A Review [article]

Boran Zhou, Xiaofeng Yang, Tian Liu
2020 arXiv   pre-print
Quantitative ultrasound (QUS) imaging is a reliable, fast and inexpensive technique to extract physically descriptive parameters for assessing pathologies.  ...  Yap et al. used the fully convolutional networks (FCNs) for semantic segmentation of breast lesions on BUS images [141] .  ...  Dezaki et al. developed a deep residual neural network (ResNet) combined with RNN for automatic recognition of cardiac cycle phase [78] .  ... 
arXiv:2003.11658v1 fatcat:iujuh7gra5ax7od2gxoo6yrbpe

Recent advances and clinical applications of deep learning in medical image analysis [article]

Xuxin Chen, Ximin Wang, Ke Zhang, Roy Zhang, Kar-Ming Fung, Theresa C. Thai, Kathleen Moore, Robert S. Mannel, Hong Liu, Bin Zheng, Yuchen Qiu
2021 arXiv   pre-print
Deep learning has received extensive research interest in developing new medical image processing algorithms, and deep learning based models have been remarkably successful in a variety of medical imaging  ...  scenarios, including classification, segmentation, detection, and image registration.  ...  allowing for training very deep networks.  ... 
arXiv:2105.13381v2 fatcat:2k342a6rhjaavpoa2qoqxhg5rq
« Previous Showing results 1 — 15 out of 145 results