Filters








4,632 Hits in 5.3 sec

Convolutional Random Walk Networks for Semantic Image Segmentation [article]

Gedas Bertasius, Lorenzo Torresani, Stella X. Yu, Jianbo Shi
2017 arXiv   pre-print
Most current semantic segmentation methods rely on fully convolutional networks (FCNs).  ...  In this work we introduce a simple, yet effective Convolutional Random Walk Network (RWN) that addresses the issues of poor boundary localization and spatially fragmented predictions with very little increase  ...  Introduction Fully convolutional networks (FCNs) were first introduced in [20] where they were shown to yield significant improvements in semantic image segmentation.  ... 
arXiv:1605.07681v3 fatcat:so4dvijel5hobaghkkv3txenge

Convolutional Random Walk Networks for Semantic Image Segmentation

Gedas Bertasius, Lorenzo Torresani, Stella X. Yu, Jianbo Shi
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Most current semantic segmentation methods rely on fully convolutional networks (FCNs).  ...  In this work, we address this problem by introducing Convolutional Random Walk Networks (RWNs) that combine the strengths of FCNs and random walk based methods.  ...  The main goal of this work was to show that our proposed random walk based spatial grouping mechanism can improve the segmentation accuracy when it is applied in the deep low resolution network layers.  ... 
doi:10.1109/cvpr.2017.650 dblp:conf/cvpr/BertasiusTYS17 fatcat:dhgrbaw4mjdptokayj2e5itedy

Learning Pixel-Level Semantic Affinity with Image-Level Supervision for Weakly Supervised Semantic Segmentation

Jiwoon Ahn, Suha Kwak
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
The semantic propagation is then realized by random walk with the affinities predicted by AffinityNet.  ...  To this end, we propose a Deep Neural Network (DNN) called AffinityNet that predicts semantic affinity between a pair of adjacent image coordinates.  ...  Sparse activations in CAMs are then diffused by random walk [23] on the graph, for each class: The affinities on edges in the graph encourage random walk to propagate the activations to nearby and semantically  ... 
doi:10.1109/cvpr.2018.00523 dblp:conf/cvpr/AhnK18 fatcat:w7gy6okeyzczndelvs2dubywde

Learning Pixel-level Semantic Affinity with Image-level Supervision for Weakly Supervised Semantic Segmentation [article]

Jiwoon Ahn, Suha Kwak
2018 arXiv   pre-print
The semantic propagation is then realized by random walk with the affinities predicted by AffinityNet.  ...  To this end, we propose a Deep Neural Network (DNN) called AffinityNet that predicts semantic affinity between a pair of adjacent image coordinates.  ...  Sparse activations in CAMs are then diffused by random walk [23] on the graph, for each class: The affinities on edges in the graph encourage random walk to propagate the activations to nearby and semantically  ... 
arXiv:1803.10464v2 fatcat:s72y6bmfgjhpzmblpqf72mwqkq

Robust Loop Closure Detection Integrating Visual–Spatial–Semantic Information via Topological Graphs and CNN Features

Yuwei Wang, Yuanying Qiu, Peitao Cheng, Xuechao Duan
2020 Remote Sensing  
In this paper, a robust loop closure detection approach integrating visual–spatial–semantic information is proposed by employing topological graphs and convolutional neural network (CNN) features.  ...  Firstly, to reduce mismatches under different viewpoints, semantic topological graphs are introduced to encode the spatial relationships of landmarks, and random walk descriptors are employed to characterize  ...  Acknowledgments: We thank Mark Cummins for providing the City Centre, New College datasets. In addition, we are also very grateful for the Gardens Point dataset provided by QUT.  ... 
doi:10.3390/rs12233890 fatcat:mjlzmdsoqbak7f5aq6rhheqoqi

Weakly-Supervised Image Semantic Segmentation Using Graph Convolutional Networks [article]

Shun-Yi Pan, Cheng-You Lu, Shih-Po Lee, Wen-Hsiao Peng
2021 arXiv   pre-print
This work addresses weakly-supervised image semantic segmentation based on image-level class labels.  ...  One common approach to this task is to propagate the activation scores of Class Activation Maps (CAMs) using a random-walk mechanism in order to arrive at complete pseudo labels for training a semantic  ...  INTRODUCTION Image semantic segmentation aims for classifying pixels in an image into their semantic classes. Training a semantic segmentation network often requires costly pixel-wise class labels.  ... 
arXiv:2103.16762v2 fatcat:3gzvbopnqbam7oylq5gfgbb7vq

DifNet: Semantic Segmentation by Diffusion Networks [article]

Peng Jiang and Fanglin Gu and Yunhai Wang and Changhe Tu and Baoquan Chen
2018 arXiv   pre-print
We use one branch network for one sub-task each, and apply a cascade of random walks base on hierarchical semantics to approximate a complex diffusion process which propagates seed information to the whole  ...  Deep Neural Networks (DNNs) have recently shown state of the art performance on semantic segmentation tasks, however, they still suffer from problems of poor boundary localization and spatial fragmented  ...  Introduction Semantic Segmentation who aims to give dense label predictions for pixels in an image is one of the fundamental topics in computer vision.  ... 
arXiv:1805.08015v4 fatcat:5umvxcwr2rgtnmv3dr2didzelu

GLOBAL MESSAGE PASSING IN NETWORKS VIA TASK-DRIVEN RANDOM WALKS FOR SEMANTIC SEGMENTATION OF REMOTE SENSING IMAGES

L. Mou, Y. Hua, P. Jin, X. X. Zhu
2020 ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences  
The capability of globally modeling and reasoning about relations between image regions is crucial for complex scene understanding tasks such as semantic segmentation.  ...  Most current semantic segmentation methods fall back on deep convolutional neural networks (CNNs), while their use of convolutions with local receptive fields is typically inefficient at capturing long-range  ...  Learning Random Walk Sampling We consider learning a random walk with t steps operating across grids.  ... 
doi:10.5194/isprs-annals-v-2-2020-533-2020 fatcat:55w7y77y5zbwplrnmdx4gsr43a

Fine-Grained Semantic Segmentation of Motion Capture Data using Dilated Temporal Fully-Convolutional Networks [article]

Noshaba Cheema, Somayeh Hosseini, Janis Sprenger, Erik Herrmann, Han Du, Klaus Fischer, Philipp Slusallek
2019 arXiv   pre-print
It first transforms a motion capture sequence into a "motion image" and applies a convolutional neural network for image segmentation.  ...  Our model outperforms two state-of-the-art models for action segmentation, as well as a popular network for sequence modeling.  ...  I.e. the distance from foot to floor works for segmenting walking actions but to segment a picking action.  ... 
arXiv:1903.00695v1 fatcat:umxzzd22lza7zkbdtc6774lofe

MeshWalker: Deep Mesh Understanding by Random Walks [article]

Alon Lahav, Ayellet Tal
2020 arXiv   pre-print
We show that our approach achieves state-of-the-art results for two fundamental shape analysis tasks: shape classification and semantic segmentation.  ...  The few attempts to answer this question propose to adapt convolutions & pooling to suit Convolutional Neural Networks (CNNs).  ...  For the given mesh we generate multiple random walks. ese walks are run through the trained network.  ... 
arXiv:2006.05353v3 fatcat:5igmnurg25g3lhunsultzriqui

Survey on Semantic Segmentation using Deep Learning Techniques

Fahad Lateef, Yassine Ruichek
2019 Neurocomputing  
Semantic segmentation is a challenging task in computer vision systems.  ...  For this reason, we propose to survey these methods by, first categorizing them into ten different classes according to the common concepts underlying their architectures.  ...  ACKNOWLEDGMENT The authors express their gratitude to University Technology Belfort-Montbeliard and Higher Education Commission of Pakistan for providing the support and necessary requirement for completion  ... 
doi:10.1016/j.neucom.2019.02.003 fatcat:aelsfl7unvdw5j2rtyqhtgqrsm

PCAMs: Weakly Supervised Semantic Segmentation Using Point Supervision [article]

R. Austin McEver, B.S. Manjunath
2020 arXiv   pre-print
This paper presents a novel procedure for producing semantic segmentation from images given some point level annotations.  ...  This method includes point annotations in the training of a convolutional neural network (CNN) for producing improved localization and class activation maps.  ...  semantic segmentation network.  ... 
arXiv:2007.05615v1 fatcat:qa3locldtjcajgicvnt4ttnfii

Weakly Supervised Semantic Segmentation of Satellite Images [article]

Adrien Nivaggioli, Hicham Randrianarivo
2019 arXiv   pre-print
When one wants to train a neural network to perform semantic segmentation, creating pixel-level annotations for each of the images in the database is a tedious task.  ...  Also, the quality of semantic segmentation performed directly by the AffinityNet and the Random Walk is close to the one of the best fully-supervised approaches.  ...  The method can be split into 4 different parts as shown in fig. 1 : Classification Network, Affinity Network, Random Walk and Segmentation Network. A.  ... 
arXiv:1904.03983v1 fatcat:77z5hbyu6vcxvor5ziwgnb2t6u

A Brief Survey on Weakly Supervised Semantic Segmentation

Youssef Ouassit, Soufiane Ardchir, Mohammed Yassine El Ghoumari, Mohamed Azouazi
2022 International Journal of Online and Biomedical Engineering (iJOE)  
Semantic Segmentation is the process of assigning a label to every pixel in the image that share same semantic properties and stays a challenging task in computer vision.  ...  In recent years, and due to the large availability of training data the performance of semantic segmentation has been greatly improved by using deep learning techniques.  ...  ] Transposed Layer SegNet 2015 SegNet: A Deep Convolutional Encoder-Decoder Archi-tecture for Image Segmentation [10] Encoder/Decoder UNet 2015 U-Net: Convolutional Networks for Biomedical Image Segmentation  ... 
doi:10.3991/ijoe.v18i10.31531 fatcat:6klflaiecrdgrizzlpgybimt6q

Affinity LCFCN: Learning to Segment Fish with Weak Supervision [article]

Issam Laradji, Alzayat Saleh, Pau Rodriguez, Derek Nowrouzezahrai, Mostafa Rahimi Azghadi, David Vazquez
2020 arXiv   pre-print
We aggregate these two outputs using a random walk to obtain the final, refined per-pixel segmentation output.  ...  Our approach uses a fully convolutional neural network with one branch that outputs per-pixel scores and another that outputs an affinity matrix.  ...  The random walk encourages neighboring pixels to have similar probabilities based on their semantic similarities.  ... 
arXiv:2011.03149v1 fatcat:kpgc47mz5vbmnacwce6ppwotgi
« Previous Showing results 1 — 15 out of 4,632 results