Filters








2,753 Hits in 5.6 sec

Virtual to Real adaptation of Pedestrian Detectors [article]

Luca Ciampi, Nicola Messina, Fabrizio Falchi, Claudio Gennaro, Giuseppe Amato
2020 arXiv   pre-print
To this end, we introduce ViPeD (Virtual Pedestrian Dataset), a new synthetically generated set of images collected with the highly photo-realistic graphical engine of the video game GTA V - Grand Theft  ...  Experiments show that the network trained with ViPeD can generalize over unseen real-world scenarios better than the detector trained over real-world data, exploiting the variety of our synthetic dataset  ...  In our opinion, the result of this work opens new perspectives to address the scalability of pedestrian and object detection methods for large physical systems with limited supervisory resources.  ... 
arXiv:2001.03032v2 fatcat:237q7npudbeyveuophbuyqloya

Virtual to Real Adaptation of Pedestrian Detectors

Luca Ciampi, Nicola Messina, Fabrizio Falchi, Claudio Gennaro, Giuseppe Amato
2020 Sensors  
To this end, we introduce ViPeD (Virtual Pedestrian Dataset), a new synthetically generated set of images collected with the highly photo-realistic graphical engine of the video game GTA V (Grand Theft  ...  Experiments show that the network trained with ViPeD can generalize over unseen real-world scenarios better than the detector trained over real-world data, exploiting the variety of our synthetic dataset  ...  On the other hand, they will also have the possibility of further specializing the detector to work over new added real-world scenarios using our two domain adaptation techniques, obtaining an additional  ... 
doi:10.3390/s20185250 pmid:32937977 fatcat:m275dyk4cbc65hg2xxh4qrl5ym

Adapting a Pedestrian Detector by Boosting LDA Exemplar Classifiers

Jiaolong Xu, David Vazquez, Sebastian Ramos, Antonio M. Lopez, Daniel Ponsa
2013 2013 IEEE Conference on Computer Vision and Pattern Recognition Workshops  
Therefore, in order to avoid this problem, it is required to adapt the detector trained with synthetic data to operate in the real-world scenario.  ...  Training vision-based pedestrian detectors using synthetic datasets (virtual world) is a useful technique to collect automatically the training examples with their pixelwise ground truth.  ...  Boosting Exemplar Classifiers for Domain Adaptation We aim at adapting a pedestrian detector trained in a source virtual world to operate in a target real world.  ... 
doi:10.1109/cvprw.2013.104 dblp:conf/cvpr/XuVRLP13 fatcat:c22ra42svrcwboj27desdmo2ia

Virtual and Real World Adaptation for Pedestrian Detection

David Vazquez, Antonio M. Lopez, Javier Marin, Daniel Ponsa, David Geronimo
2014 IEEE Transactions on Pattern Analysis and Machine Intelligence  
To the best of our knowledge, this is the first work demonstrating adaptation of virtual and real worlds for developing an object detector.  ...  with the many examples of the source domain (virtual world) in order to train a domain adapted pedestrian classifier that will operate in the target domain.  ...  To the best of our knowledge, this is the first work demonstrating adaptation of virtual and real worlds for developing an appearance-based object detector.  ... 
doi:10.1109/tpami.2013.163 pmid:26353201 fatcat:igb6ie7fcfcblldqxyiu4ymepm

Scene-Specific Pedestrian Detection Based on Parallel Vision [article]

Wenwen Zhang, Kunfeng Wang, Hua Qu, Jihong Zhao, Fei-Yue Wang
2017 arXiv   pre-print
to appear to solve this problem of lacking labeled data and the results show that data from virtual world is helpful to adapt generic pedestrian detectors to specific scenes.  ...  In order to make the generic scene pedestrian detectors work well in specific scenes, the labeled data from specific scenes are needed to adapt the models to the specific scenes.  ...  Our work is also to adapting a generic pedestrian detector to a specific scene without real data, similar to their work [8] , but we are not train many detectors in each possible position.  ... 
arXiv:1712.08745v1 fatcat:y6f3iuo5hbh3vmbq5ftf5sczte

Weakly Supervised Automatic Annotation of Pedestrian Bounding Boxes

David Vazquez, Jiaolong Xu, Sebastian Ramos, Antonio M. Lopez, Daniel Ponsa
2013 2013 IEEE Conference on Computer Vision and Pattern Recognition Workshops  
Among the components of a pedestrian detector, its trained pedestrian classifier is crucial for achieving the desired performance.  ...  Accordingly, in this paper we assess an strategy to collect samples from the real world and retrain with them, thus avoiding the dataset shift, but in such a way that no BBs of real-world pedestrians have  ...  However, rather than devising domain adaptation procedures, we propose to use the virtual-world data for developing a pedestrian classifier to be used for collecting pedestrian detections from real-world  ... 
doi:10.1109/cvprw.2013.107 dblp:conf/cvpr/VazquezXRLP13 fatcat:5b657dmquvbcpdhyg6wyuk4r6y

Learning scene-specific pedestrian detectors without real data

Hironori Hattori, Vishnu Naresh Boddeti, Kris Kitani, Takeo Kanade
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
We consider the problem of designing a scene-specific pedestrian detector in a scenario where we have zero instances of real pedestrian data (i.e., no labeled real data or unsupervised real data).  ...  The key idea of our approach is to infer the potential appearance of pedestrians using geometric scene data and a customizable database of virtual simulations of pedestrian motion.  ...  Our work is different in that we do not use real data from the scene to adapt our detector. Scene-Specific Domain Adaptation.  ... 
doi:10.1109/cvpr.2015.7299006 dblp:conf/cvpr/HattoriBKK15 fatcat:akli7fz3lvbvzanozub5kn67ku

Domain Adaptation of Deformable Part-Based Models

Jiaolong Xu, Sebastian Ramos, David Vazquez, Antonio M. Lopez
2014 IEEE Transactions on Pattern Analysis and Machine Intelligence  
Two types of adaptation tasks are assessed: from both synthetic pedestrians and general persons (PASCAL VOC) to pedestrians imaged from an on-board camera.  ...  Neither A-SSVM nor SA-SSVM needs to revisit the source-domain training data to perform the adaptation. Rather, a low number of target-domain training examples (e.g., pedestrians) are used.  ...  Bagdanov from the Computer Vision Center for helping us to improve paper readability.  ... 
doi:10.1109/tpami.2014.2327973 pmid:26353145 fatcat:5c6tlp4vmnf3vpwe4elewdmzay

Learning appearance in virtual scenarios for pedestrian detection

Javier Marin, David Vazquez, David Geronimo, Antonio M. Lopez
2010 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition  
The comparison reveals that, although virtual samples were not specially selected, both virtual and real based training give rise to classifiers of similar performance.  ...  Detecting pedestrians in images is a key functionality to avoid vehicle-to-pedestrian collisions.  ...  We assess the similarity of virtual and real worldbased training, both in terms of the performance of the corresponding pedestrian detectors and the matching of the specific detection results.  ... 
doi:10.1109/cvpr.2010.5540218 dblp:conf/cvpr/MarinVGL10 fatcat:a7e3ocrplbgejbdmm6g6niuvzq

Partially fake it till you make it: mixing real and fake thermal images for improved object detection [article]

Francesco Bongini, Lorenzo Berlincioni, Marco Bertini, Alberto Del Bimbo
2021 arXiv   pre-print
to best combine our proposed augmentation with these other techniques.Experimental results demonstrate the effectiveness of our approach, and our single-modality detector achieves state-of-the-art results  ...  full realistic synthetic scenes is extremely cumbersome and expensive due to the difficulty in modeling the thermal properties of the materials of the scene.  ...  Fig. 3 shows an example of the process: a real scene from FLIR dataset is used to composite and animate 3D objects, creating a virtual video sequence.  ... 
arXiv:2106.13603v1 fatcat:f2skg4qvlzg6rizopoae2slypm

From Virtual to Real World Visual Perception using Domain Adaptation -- The DPM as Example [article]

Antonio M. Lopez, Jiaolong Xu, Jose L. Gomez, David Vazquez, German Ros
2016 arXiv   pre-print
However, since the models learned with virtual data must operate in the real world, we still need to perform domain adaptation (DA).  ...  In this chapter we revisit the DA of a deformable part-based model (DPM) as an exemplifying case of virtual- to-real-world DA.  ...  ing case of virtual- to real-world DA.  ... 
arXiv:1612.09134v1 fatcat:ky7jz3nvh5gkjaxe5npoztyooi

Augmentation of virtual agents in real crowd videos

Yalım Doğan, Serkan Demirci, Uğur Güdükbay, Hamdi Dibeklioğlu
2018 Signal, Image and Video Processing  
Augmenting virtual agents in real crowd videos is an important task for different applications from simulations of social environments to modeling abnormalities in crowd behavior.  ...  We utilize pedestrian detection and tracking algorithms to automatically locate the pedestrians in video frames and project them into our simulated environment, where the navigable area of the simulated  ...  Fig. 2 2 Pedestrian detection and background subtraction: a the output of the pedestrian detector, b the output of the background subtractor, c the output of the pedestrian detector with background subtraction  ... 
doi:10.1007/s11760-018-1392-8 fatcat:rii2fgemnjfqtgcvukup4ejhi4

Expecting the Unexpected: Training Detectors for Unusual Pedestrians with Adversarial Imposters [article]

Shiyu Huang, Deva Ramanan
2017 arXiv   pre-print
To analyze this problem, we have collected a novel annotated dataset of dangerous scenarios called the Precarious Pedestrian dataset.  ...  challenge for real-world deployment.  ...  (b) 3D models that we use in this project. ple pipeline for adapting detectors from synthetic data to the world of real images.  ... 
arXiv:1703.06283v2 fatcat:t3enibfxavepjozn7ss54ztnla

Expecting the Unexpected: Training Detectors for Unusual Pedestrians with Adversarial Imposters

Shiyu Huang, Deva Ramanan
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
To analyze this problem, we have collected a novel annotated dataset of dangerous scenarios called the Precarious Pedestrian dataset.  ...  As autonomous vehicles become an every-day reality, high-accuracy pedestrian detection is of paramount practical importance.  ...  Large synthetic datasets can be used to bootstrap detectors and then adapted to real data by moving to the target domain distribution.  ... 
doi:10.1109/cvpr.2017.496 dblp:conf/cvpr/HuangR17 fatcat:kgmfcfdknzac3iwgi7widajdsq

Learning a multiview part-based model in virtual world for pedestrian detection

Jiaolong Xu, David Vazquez, Antonio M. Lopez, Javier Marin, Daniel Ponsa
2013 2013 IEEE Intelligent Vehicles Symposium (IV)  
The method is efficient as: (i) the part detectors are trained with precisely extracted virtual examples, thus no latent learning is needed, (ii) the multiview pedestrian detector enhances the performance  ...  In this paper, we propose to train a multiview deformable part-based model with automatically generated part examples from virtual-world data.  ...  The developed pedestrian detector, based on a holistic HOG/Linear-SVM pedestrian classifier, showed a performance comparable to analogous detectors obtained from real-world manually labelled data.  ... 
doi:10.1109/ivs.2013.6629512 dblp:conf/ivs/XuVLMP13 fatcat:5skx3sik45athek27f4diykz5i
« Previous Showing results 1 — 15 out of 2,753 results