Filters








3,119 Hits in 5.7 sec

Is object localization for free? - Weakly-supervised learning with convolutional neural networks

Maxime Oquab, Leon Bottou, Ivan Laptev, Josef Sivic
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
We describe a weakly supervised convolutional neural network (CNN) for object classification that relies only on image-level labels, yet can learn from cluttered scenes containing multiple objects.  ...  Detailed image annotation, e.g. by object bounding boxes, however, is both expensive and often subjective.  ...  So, is object localization with convolutional neural networks for free?  ... 
doi:10.1109/cvpr.2015.7298668 dblp:conf/cvpr/OquabBLS15 fatcat:ivhq7y7465gbhlj6ouaqw5sk3i

Learning Saliency-Free Model with Generic Features for Weakly-Supervised Semantic Segmentation

Wenfeng Luo, Meng Yang
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Current weakly-supervised semantic segmentation methods often estimate initial supervision from class activation maps (CAM), which produce sparse discriminative object seeds and rely on image saliency  ...  Experiments on the PASCAL VOC 2012 dataset show that the proposed saliency-free method outperforms the previous approaches under the same weakly-supervised setting and achieves superior segmentation results  ...  , the Guangzhou Science and Technology Program (Grant no. 201804010288), and the Fundamental Research Funds for the Central Universities (Grant no.18lgzd15).  ... 
doi:10.1609/aaai.v34i07.6842 fatcat:tbb25v7tbreabmmsre2r2yoccq

Detector-Free Weakly Supervised Group Activity Recognition [article]

Dongkeun Kim, Jinsung Lee, Minsu Cho, Suha Kwak
2022 arXiv   pre-print
Motivated by this, we propose a novel model for group activity recognition that depends neither on bounding box labels nor on object detector.  ...  Existing models for this task are often impractical in that they demand ground-truth bounding box labels of actors even in testing or rely on off-the-shelf object detectors.  ...  It bypasses explicit object detection by drawing attention on entities involved in a group activity through a Transformer encoder [49] placed on top of a convolutional neural network (CNN) backbone.  ... 
arXiv:2204.02139v1 fatcat:64gp6rrlfrdo3nptrpess65vya

Pseudo-label-free Weakly Supervised Semantic Segmentation Using Image Masking

Sangtae Kim, Luong Trung Nguyen, Kyuhong Shim, Junhan Kim, Byonghyo Shim
2022 IEEE Access  
Weakly-supervised semantic segmentation (WSSS) aims to train a semantic segmentation network using weak labels.  ...  Key idea of the proposed approach is to train the segmentation network such that the object erased by the segmentation map is not detected by the classification network.  ...  The generated pseudo-label depicting the reliable object regions is used as a supervision for the semantic segmentation network.  ... 
doi:10.1109/access.2022.3149587 fatcat:4xloz7mbhff6xpheqcozdmihh4

OrigamiNet: Weakly-Supervised, Segmentation-Free, One-Step, Full Page Text Recognition by learning to unfold [article]

Mohamed Yousef, Tom E. Bishop
2020 arXiv   pre-print
We propose a novel and simple neural network module, termed OrigamiNet, that can augment any CTC-trained, fully convolutional single line text recognizer, to convert it into a multi-line version by providing  ...  Text recognition is a major computer vision task with a big set of associated challenges. One of those traditional challenges is the coupled nature of text recognition and segmentation.  ...  We then concluded with a set of interpretability experiments to investigate what the model actually learns and demonstrated its implicit ability to localize characters on each line.  ... 
arXiv:2006.07491v1 fatcat:egcvahk6xrf6xd3tuinddyhyfe

OrigamiNet: Weakly-Supervised, Segmentation-Free, One-Step, Full Page Text Recognition by learning to unfold

Mohamed Yousef, Tom E. Bishop
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
We propose a novel and simple neural network module, termed OrigamiNet, that can augment any CTC-trained, fully convolutional single line text recognizer, to convert it into a multi-line version by providing  ...  Text recognition is a major computer vision task with a big set of associated challenges. One of those traditional challenges is the coupled nature of text recognition and segmentation.  ...  In this work, we present a simple and novel neural network sub-module, termed OrigamiNet, that can be added to any existing convolutional neural network (CNN) text-line recognizer to convert it to a full  ... 
doi:10.1109/cvpr42600.2020.01472 dblp:conf/cvpr/YousefB20 fatcat:fgr3tluvg5hgbbolcujar7jsca

Minimizing Supervision for Free-space Segmentation [article]

Satoshi Tsutsui, Tommi Kerola, Shunta Saito, David J. Crandall
2018 arXiv   pre-print
Although weakly supervised segmentation addresses this issue, most methods are not designed for free-space.  ...  In this paper, we observe that homogeneous texture and location are two key characteristics of free-space, and develop a novel, practical framework for free-space segmentation with minimal human supervision  ...  We would also like to thank Masaki Saito and Richard Calland for helpful discussions.  ... 
arXiv:1711.05998v3 fatcat:hnnjaeuevffmjli7dfqlkdcnya

DeformNet: Free-Form Deformation Network for 3D Shape Reconstruction from a Single Image [article]

Andrey Kurenkov, Jingwei Ji, Animesh Garg, Viraj Mehta, JunYoung Gwak, Christopher Choy, Silvio Savarese
2017 arXiv   pre-print
We evaluate our approach on the ShapeNet dataset and show that - (a) the Free-Form Deformation layer is a powerful new building block for Deep Learning models that manipulate 3D data (b) DeformNet uses  ...  We introduce a new differentiable layer for 3D data deformation and use it in DeformNet to learn a model for 3D reconstruction-through-deformation.  ...  There are also representative advances in unsupervised/weakly supervised 3D learning for reconstruction.  ... 
arXiv:1708.04672v1 fatcat:tdbjw6rn5bch3kuidapivqgsd4

Minimizing Supervision for Free-Space Segmentation

Satoshi Tsutsui, Tommi Kerola, Shunta Saito, David J. Crandall
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)  
Although weakly supervised segmentation addresses this issue, most methods are not designed for free-space.  ...  In this paper, we observe that homogeneous texture and location are two key characteristics of free-space, and develop a novel, practical framework for free-space segmentation with minimal human supervision  ...  We would also like to thank Masaki Saito and Richard Calland for helpful discussions.  ... 
doi:10.1109/cvprw.2018.00145 dblp:conf/cvpr/TsutsuiKSC18 fatcat:wt3rfaphtngztb3y4uzpxbaw7a

Attentive Prototypes for Source-free Unsupervised Domain Adaptive 3D Object Detection [article]

Deepti Hegde, Vishal M. Patel
2021 arXiv   pre-print
We propose a single-frame approach for source-free, unsupervised domain adaptation of lidar-based 3D object detectors that uses class prototypes to mitigate the effect pseudo-label noise.  ...  3D object detection networks tend to be biased towards the data they are trained on.  ...  Originally calculated with handcrafted features [13] , recent approaches use convolutional neural networks for feature extraction.  ... 
arXiv:2111.15656v2 fatcat:6lzmcwufyzfxtni53swxkiuohi

Relaxing from Vocabulary: Robust Weakly-Supervised Deep Learning for Vocabulary-Free Image Tagging

Jianlong Fu, Yue Wu, Tao Mei, Jinqiao Wang, Hanqing Lu, Yong Rui
2015 2015 IEEE International Conference on Computer Vision (ICCV)  
In this paper, we propose a weakly-supervised deep learning model which can be trained from the readily available Web images to relax the dependence on human labors and scale up to arbitrary tags (categories  ...  The discrepancy is finally transformed into the objective function to give relevance feedback to back propagation.  ...  In this paper, we propose a robust weakly-supervised deep learning network with the noisy Web training data for image tagging.  ... 
doi:10.1109/iccv.2015.230 dblp:conf/iccv/FuWMWLR15 fatcat:7w3bpjecorarnbweam4ddwszcu

A Likelihood-Free Inference Framework for Population Genetic Data using Exchangeable Neural Networks [article]

Jeffrey Chan, Valerio Perrone, Jeffrey P Spence, Paul A Jenkins, Sara Mathieson, Yun S Song
2018 bioRxiv   pre-print
This is achieved by means of a novel Bayesian likelihood-free inference framework, where a permutation-invariant convolutional neural network learns the inverse functional relationship from the data to  ...  In this paper, we learn the first exchangeable feature representation for population genetic data to work directly with genotype data.  ...  YSS is a Chan Zuckerberg Biohub  ... 
doi:10.1101/267211 fatcat:hofwzezsqfaybjrztesvjqe3ti

A Likelihood-Free Inference Framework for Population Genetic Data using Exchangeable Neural Networks [article]

Jeffrey Chan, Valerio Perrone, Jeffrey P. Spence, Paul A. Jenkins, Sara Mathieson, Yun S. Song
2018 arXiv   pre-print
In this work, we develop an exchangeable neural network that performs summary statistic-free, likelihood-free inference.  ...  is intractable as it requires integrating over a set of correlated, extremely high-dimensional latent variables.  ...  For example, with 4 supervised learning, overfitting is a result of large estimation error.  ... 
arXiv:1802.06153v2 fatcat:uv63a54qrfghzgotg5q3l2cv2a

Parameter-Free Spatial Attention Network for Person Re-Identification [article]

Haoran Wang, Yue Fan, Zexin Wang, Licheng Jiao, Bernt Schiele
2018 arXiv   pre-print
While GAP helps the convolution neural network to attend to the most discriminative features of an object, it may suffer if that information is missing e.g. due to camera viewpoint changes.  ...  Global average pooling (GAP) allows to localize discriminative information for recognition [40].  ...  Related Work Weakly Supervised Learning.  ... 
arXiv:1811.12150v1 fatcat:6cjfafo2mff6fpghsufylfgkfe

Multi-Graph Transformer for Free-Hand Sketch Recognition [article]

Peng Xu, Chaitanya K. Joshi, Xavier Bresson
2021 arXiv   pre-print
Existing techniques have focused on exploiting either the static nature of sketches with Convolutional Neural Networks (CNNs) or the temporal sequential property with Recurrent Neural Networks (RNNs).  ...  We design a novel Graph Neural Network (GNN), the Multi-Graph Transformer (MGT), for learning representations of sketches from multiple graphs which simultaneously capture global and local geometric stroke  ...  Additionally, for the graph neural network (GNN) community, we hope that MGT helps free-hand sketch become a new test-bed for GNNs. VI.  ... 
arXiv:1912.11258v3 fatcat:pxurw2kxbffzdn4esi2ydkejtm
« Previous Showing results 1 — 15 out of 3,119 results