Publicacions CVC -- Query Results

	Publicacions CVC Home \| Show All \| Simple Search \| Advanced Search \| Add Record \| Import	Login Quick Search: Field: contains: ...
	3151–3165 of 3413 records found matching your query (RSS):

Search & Display Options

Select All Deselect All

[201–210] << 211 212 213 214 215 216 217 218 219 220 >> [221–228]

List View

Citations

Details

	Records
	Author	Emanuele Vivoli; Ali Furkan Biten; Andres Mafla; Dimosthenis Karatzas; Lluis Gomez
	Title	MUST-VQA: MUltilingual Scene-text VQA			Type	Conference Article
	Year	2022	Publication	Proceedings European Conference on Computer Vision Workshops	Abbreviated Journal
	Volume	13804	Issue		Pages	345–358
	Keywords	Visual question answering; Scene text; Translation robustness; Multilingual models; Zero-shot transfer; Power of language models
	Abstract	In this paper, we present a framework for Multilingual Scene Text Visual Question Answering that deals with new languages in a zero-shot fashion. Specifically, we consider the task of Scene Text Visual Question Answering (STVQA) in which the question can be asked in different languages and it is not necessarily aligned to the scene text language. Thus, we first introduce a natural step towards a more generalized version of STVQA: MUST-VQA. Accounting for this, we discuss two evaluation scenarios in the constrained setting, namely IID and zero-shot and we demonstrate that the models can perform on a par on a zero-shot setting. We further provide extensive experimentation and show the effectiveness of adapting multilingual language models into STVQA tasks.
	Address	Tel-Aviv; Israel; October 2022
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title	LNCS
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference	ECCVW
	Notes	DAG; 302.105; 600.155; 611.002			Approved	no
	Call Number	Admin @ si @ VBM2022			Serial	3770
Permanent link to this record



	Author	Sergi Garcia Bordils; Andres Mafla; Ali Furkan Biten; Oren Nuriel; Aviad Aberdam; Shai Mazor; Ron Litman; Dimosthenis Karatzas
	Title	Out-of-Vocabulary Challenge Report			Type	Conference Article
	Year	2022	Publication	Proceedings European Conference on Computer Vision Workshops	Abbreviated Journal
	Volume	13804	Issue		Pages	359–375
	Keywords
	Abstract	This paper presents final results of the Out-Of-Vocabulary 2022 (OOV) challenge. The OOV contest introduces an important aspect that is not commonly studied by Optical Character Recognition (OCR) models, namely, the recognition of unseen scene text instances at training time. The competition compiles a collection of public scene text datasets comprising of 326,385 images with 4,864,405 scene text instances, thus covering a wide range of data distributions. A new and independent validation and test set is formed with scene text instances that are out of vocabulary at training time. The competition was structured in two tasks, end-to-end and cropped scene text recognition respectively. A thorough analysis of results from baselines and different participants is presented. Interestingly, current state-of-the-art models show a significant performance gap under the newly studied setting. We conclude that the OOV dataset proposed in this challenge will be an essential area to be explored in order to develop scene text models that achieve more robust and generalized predictions.
	Address	Tel-Aviv; Israel; October 2022
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title	LNCS
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference	ECCVW
	Notes	DAG; 600.155; 302.105; 611.002			Approved	no
	Call Number	Admin @ si @ GMB2022			Serial	3771
Permanent link to this record



	Author	Patricia Suarez; Angel Sappa; Dario Carpio; Henry Velesaca; Francisca Burgos; Patricia Urdiales
	Title	Deep Learning Based Shrimp Classification			Type	Conference Article
	Year	2022	Publication	17th International Symposium on Visual Computing	Abbreviated Journal
	Volume	13598	Issue		Pages	36–45
	Keywords	Pigmentation; Color space; Light weight network
	Abstract	This work proposes a novel approach based on deep learning to address the classification of shrimp (Pennaeus vannamei) into two classes, according to their level of pigmentation accepted by shrimp commerce. The main goal of this actual study is to support the shrimp industry in terms of price and process. An efficient CNN architecture is proposed to perform image classification through a program that could be set other in mobile devices or in fixed support in the shrimp supply chain. The proposed approach is a lightweight model that uses HSV color space shrimp images. A simple pipeline shows the most important stages performed to determine a pattern that identifies the class to which they belong based on their pigmentation. For the experiments, a database acquired with mobile devices of various brands and models has been used to capture images of shrimp. The results obtained with the images in the RGB and HSV color space allow for testing the effectiveness of the proposed model.
	Address
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference	ISVC
	Notes	MSIAU; no proj			Approved	no
	Call Number	Admin @ si @ SAC2022			Serial	3772
Permanent link to this record



	Author	Henry Velesaca; Patricia Suarez; Angel Sappa; Dario Carpio; Rafael E. Rivadeneira; Angel Sanchez
	Title	Review on Common Techniques for Urban Environment Video Analytics			Type	Conference Article
	Year	2022	Publication	Anais do III Workshop Brasileiro de Cidades Inteligentes	Abbreviated Journal
	Volume		Issue		Pages	107-118
	Keywords	Video Analytics; Review; Urban Environments; Smart Cities
	Abstract	This work compiles the different computer vision-based approaches from the state-of-the-art intended for video analytics in urban environments. The manuscript groups the different approaches according to the typical modules present in video analysis, including image preprocessing, object detection, classification, and tracking. This proposed pipeline serves as a basic guide to representing these most representative approaches in this topic of video analysis that will be addressed in this work. Furthermore, the manuscript is not intended to be an exhaustive review of the most advanced approaches, but only a list of common techniques proposed to address recurring problems in this field.
	Address
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference	WBCI
	Notes	MSIAU; 601.349			Approved	no
	Call Number	Admin @ si @ VSS2022			Serial	3773
Permanent link to this record



	Author	Aneesh Rangnekar; Zachary Mulhollan; Anthony Vodacek; Matthew Hoffman; Angel Sappa; Erik Blasch; Jun Yu; Liwen Zhang; Shenshen Du; Hao Chang; Keda Lu; Zhong Zhang; Fang Gao; Ye Yu; Feng Shuang; Lei Wang; Qiang Ling; Pranjay Shyam; Kuk-Jin Yoon; Kyung-Soo Kim
	Title	Semi-Supervised Hyperspectral Object Detection Challenge Results – PBVS 2022			Type	Conference Article
	Year	2022	Publication	IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)	Abbreviated Journal
	Volume		Issue		Pages	390-398
	Keywords	Training; Computer visio; Conferences; Training data; Object detection; Semisupervised learning; Transformers
	Abstract	This paper summarizes the top contributions to the first semi-supervised hyperspectral object detection (SSHOD) challenge, which was organized as a part of the Perception Beyond the Visible Spectrum (PBVS) 2022 workshop at the Computer Vision and Pattern Recognition (CVPR) conference. The SSHODC challenge is a first-of-its-kind hyperspectral dataset with temporally contiguous frames collected from a university rooftop observing a 4-way vehicle intersection over a period of three days. The dataset contains a total of 2890 frames, captured at an average resolution of 1600 × 192 pixels, with 51 hyperspectral bands from 400nm to 900nm. SSHOD challenge uses 989 images as the training set, 605 images as validation set and 1296 images as the evaluation (test) set. Each set was acquired on a different day to maximize the variance in weather conditions. Labels are provided for 10% of the annotated data, hence formulating a semi-supervised learning task for the participants which is evaluated in terms of average precision over the entire set of classes, as well as individual moving object classes: namely vehicle, bus and bike. The challenge received participation registration from 38 individuals, with 8 participating in the validation phase and 3 participating in the test phase. This paper describes the dataset acquisition, with challenge formulation, proposed methods and qualitative and quantitative results.
	Address	New Orleans; USA; June 2022
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference	CVPRW
	Notes	MSIAU; no menciona			Approved	no
	Call Number	Admin @ si @ RMV2022			Serial	3774
Permanent link to this record



	Author	Rafael E. Rivadeneira; Angel Sappa; Boris X. Vintimilla; Jin Kim; Dogun Kim; Zhihao Li; Yingchun Jian; Bo Yan; Leilei Cao; Fengliang Qi; Hongbin Wang Rongyuan Wu; Lingchen Sun; Yongqiang Zhao; Lin Li; Kai Wang; Yicheng Wang; Xuanming Zhang; Huiyuan Wei; Chonghua Lv; Qigong Sun; Xiaolin Tian; Zhuang Jia; Jiakui Hu; Chenyang Wang; Zhiwei Zhong; Xianming Liu; Junjun Jiang
	Title	Thermal Image Super-Resolution Challenge Results – PBVS 2022			Type	Conference Article
	Year	2022	Publication	IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)	Abbreviated Journal
	Volume		Issue		Pages	418-426
	Keywords
	Abstract	This paper presents results from the third Thermal Image Super-Resolution (TISR) challenge organized in the Perception Beyond the Visible Spectrum (PBVS) 2022 workshop. The challenge uses the same thermal image dataset as the first two challenges, with 951 training images and 50 validation images at each resolution. A set of 20 images was kept aside for testing. The evaluation tasks were to measure the PSNR and SSIM between the SR image and the ground truth (HR thermal noisy image downsampled by four), and also to measure the PSNR and SSIM between the SR image and the semi-registered HR image (acquired with another camera). The results outperformed those from last year’s challenge, improving both evaluation metrics. This year, almost 100 teams participants registered for the challenge, showing the community’s interest in this hot topic.
	Address	New Orleans; USA; June 2022
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference	CVPRW
	Notes	MSIAU; no menciona			Approved	no
	Call Number	Admin @ si @ RSV2022c			Serial	3775
Permanent link to this record



	Author	Rafael E. Rivadeneira; Angel Sappa; Boris X. Vintimilla
	Title	Thermal Image Super-Resolution: A Novel Unsupervised Approach			Type	Conference Article
	Year	2022	Publication	International Joint Conference on Computer Vision, Imaging and Computer Graphics	Abbreviated Journal
	Volume	1474	Issue		Pages	495–506
	Keywords
	Abstract	This paper proposes the use of a CycleGAN architecture for thermal image super-resolution under a transfer domain strategy, where middle-resolution images from one camera are transferred to a higher resolution domain of another camera. The proposed approach is trained with a large dataset acquired using three thermal cameras at different resolutions. An unsupervised learning process is followed to train the architecture. Additional loss function is proposed trying to improve results from the state of the art approaches. Following the first thermal image super-resolution challenge (PBVS-CVPR2020) evaluations are performed. A comparison with previous works is presented showing the proposed approach reaches the best results.
	Address
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference	VISIGRAPP
	Notes	MSIAU; 600.130			Approved	no
	Call Number	Admin @ si @ RSV2022d			Serial	3776
Permanent link to this record



	Author	Marc Oliu; Sarah Adel Bargal; Stan Sclaroff; Xavier Baro; Sergio Escalera
	Title	Multi-varied Cumulative Alignment for Domain Adaptation			Type	Conference Article
	Year	2022	Publication	6th International Conference on Image Analysis and Processing	Abbreviated Journal
	Volume	13232	Issue		Pages	324–334
	Keywords	Domain Adaptation; Computer vision; Neural networks
	Abstract	Domain Adaptation methods can be classified into two basic families of approaches: non-parametric and parametric. Non-parametric approaches depend on statistical indicators such as feature covariances to minimize the domain shift. Non-parametric approaches tend to be fast to compute and require no additional parameters, but they are unable to leverage probability density functions with complex internal structures. Parametric approaches, on the other hand, use models of the probability distributions as surrogates in minimizing the domain shift, but they require additional trainable parameters to model these distributions. In this work, we propose a new statistical approach to minimizing the domain shift based on stochastically projecting and evaluating the cumulative density function in both domains. As with non-parametric approaches, there are no additional trainable parameters. As with parametric approaches, the internal structure of both domains’ probability distributions is considered, thus leveraging a higher amount of information when reducing the domain shift. Evaluation on standard datasets used for Domain Adaptation shows better performance of the proposed model compared to non-parametric approaches while being competitive with parametric ones. (Code available at: https://github.com/moliusimon/mca).
	Address	Indonesia; October 2022
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title	LNCS
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference	ICIAP
	Notes	HuPBA; no menciona			Approved	no
	Call Number	Admin @ si @ OAS2022			Serial	3777
Permanent link to this record



	Author	Ajian Liu; Chenxu Zhao; Zitong Yu; Jun Wan; Anyang Su; Xing Liu; Zichang Tan; Sergio Escalera; Junliang Xing; Yanyan Liang; Guodong Guo; Zhen Lei; Stan Z. Li; Shenshen Du
	Title	Contrastive Context-Aware Learning for 3D High-Fidelity Mask Face Presentation Attack Detection			Type	Journal Article
	Year	2022	Publication	IEEE Transactions on Information Forensics and Security	Abbreviated Journal	TIForensicSEC
	Volume	17	Issue		Pages	2497 - 2507
	Keywords
	Abstract	Face presentation attack detection (PAD) is essential to secure face recognition systems primarily from high-fidelity mask attacks. Most existing 3D mask PAD benchmarks suffer from several drawbacks: 1) a limited number of mask identities, types of sensors, and a total number of videos; 2) low-fidelity quality of facial masks. Basic deep models and remote photoplethysmography (rPPG) methods achieved acceptable performance on these benchmarks but still far from the needs of practical scenarios. To bridge the gap to real-world applications, we introduce a large-scale Hi gh- Fi delity Mask dataset, namely HiFiMask . Specifically, a total amount of 54,600 videos are recorded from 75 subjects with 225 realistic masks by 7 new kinds of sensors. Along with the dataset, we propose a novel C ontrastive C ontext-aware L earning (CCL) framework. CCL is a new training methodology for supervised PAD tasks, which is able to learn by leveraging rich contexts accurately (e.g., subjects, mask material and lighting) among pairs of live faces and high-fidelity mask attacks. Extensive experimental evaluations on HiFiMask and three additional 3D mask datasets demonstrate the effectiveness of our method. The codes and dataset will be released soon.
	Address
	Corporate Author				Thesis
	Publisher	IEEE	Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference
	Notes	HuPBA			Approved	no
	Call Number	Admin @ si @ LZY2022			Serial	3778
Permanent link to this record



	Author	Hugo Bertiche; Meysam Madadi; Sergio Escalera
	Title	Neural Cloth Simulation			Type	Journal Article
	Year	2022	Publication	ACM Transactions on Graphics	Abbreviated Journal	ACMTGraph
	Volume	41	Issue	6	Pages	1-14
	Keywords
	Abstract	We present a general framework for the garment animation problem through unsupervised deep learning inspired in physically based simulation. Existing trends in the literature already explore this possibility. Nonetheless, these approaches do not handle cloth dynamics. Here, we propose the first methodology able to learn realistic cloth dynamics unsupervisedly, and henceforth, a general formulation for neural cloth simulation. The key to achieve this is to adapt an existing optimization scheme for motion from simulation based methodologies to deep learning. Then, analyzing the nature of the problem, we devise an architecture able to automatically disentangle static and dynamic cloth subspaces by design. We will show how this improves model performance. Additionally, this opens the possibility of a novel motion augmentation technique that greatly improves generalization. Finally, we show it also allows to control the level of motion in the predictions. This is a useful, never seen before, tool for artists. We provide of detailed analysis of the problem to establish the bases of neural cloth simulation and guide future research into the specifics of this domain. ACM Transactions on GraphicsVolume 41Issue 6December 2022 Article No.: 220pp 1–
	Address	Dec 2022
	Corporate Author				Thesis
	Publisher	ACM	Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference
	Notes				Approved	no
	Call Number	Admin @ si @ BME2022b			Serial	3779
Permanent link to this record



	Author	Joakim Bruslund Haurum; Meysam Madadi; Sergio Escalera; Thomas B. Moeslund
	Title	Multi-scale hybrid vision transformer and Sinkhorn tokenizer for sewer defect classification			Type	Journal Article
	Year	2022	Publication	Automation in Construction	Abbreviated Journal	AC
	Volume	144	Issue		Pages	104614
	Keywords	Sewer Defect Classification; Vision Transformers; Sinkhorn-Knopp; Convolutional Neural Networks; Closed-Circuit Television; Sewer Inspection
	Abstract	A crucial part of image classification consists of capturing non-local spatial semantics of image content. This paper describes the multi-scale hybrid vision transformer (MSHViT), an extension of the classical convolutional neural network (CNN) backbone, for multi-label sewer defect classification. To better model spatial semantics in the images, features are aggregated at different scales non-locally through the use of a lightweight vision transformer, and a smaller set of tokens was produced through a novel Sinkhorn clustering-based tokenizer using distinct cluster centers. The proposed MSHViT and Sinkhorn tokenizer were evaluated on the Sewer-ML multi-label sewer defect classification dataset, showing consistent performance improvements of up to 2.53 percentage points.
	Address	Dec 2022
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference
	Notes	HuPBA			Approved	no
	Call Number	Admin @ si @ BME2022c			Serial	3780
Permanent link to this record



	Author	Aitor Alvarez-Gila; Joost Van de Weijer; Yaxing Wang; Estibaliz Garrote
	Title	MVMO: A Multi-Object Dataset for Wide Baseline Multi-View Semantic Segmentation			Type	Conference Article
	Year	2022	Publication	29th IEEE International Conference on Image Processing	Abbreviated Journal
	Volume		Issue		Pages
	Keywords	multi-view; cross-view; semantic segmentation; synthetic dataset
	Abstract	We present MVMO (Multi-View, Multi-Object dataset): a synthetic dataset of 116,000 scenes containing randomly placed objects of 10 distinct classes and captured from 25 camera locations in the upper hemisphere. MVMO comprises photorealistic, path-traced image renders, together with semantic segmentation ground truth for every view. Unlike existing multi-view datasets, MVMO features wide baselines between cameras and high density of objects, which lead to large disparities, heavy occlusions and view-dependent object appearance. Single view semantic segmentation is hindered by self and inter-object occlusions that could benefit from additional viewpoints. Therefore, we expect that MVMO will propel research in multi-view semantic segmentation and cross-view semantic transfer. We also provide baselines that show that new research is needed in such fields to exploit the complementary information of multi-view setups 1 .
	Address	Bordeaux; France; October2022
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference	ICIP
	Notes	LAMP			Approved	no
	Call Number	Admin @ si @ AWW2022			Serial	3781
Permanent link to this record



	Author	Nil Ballus; Bhalaji Nagarajan; Petia Radeva
	Title	Opt-SSL: An Enhanced Self-Supervised Framework for Food Recognition			Type	Conference Article
	Year	2022	Publication	10th Iberian Conference on Pattern Recognition and Image Analysis	Abbreviated Journal
	Volume	13256	Issue		Pages
	Keywords	Self-supervised; Contrastive learning; Food recognition
	Abstract	Self-supervised Learning has been showing upbeat performance in several computer vision tasks. The popular contrastive methods make use of a Siamese architecture with different loss functions. In this work, we go deeper into two very recent state of the art frameworks, namely, SimSiam and Barlow Twins. Inspired by them, we propose a new self-supervised learning method we call Opt-SSL that combines both image and feature contrasting. We validate the proposed method on the food recognition task, showing that our proposed framework enables the self-learning networks to learn better visual representations.
	Address	Aveiro; Portugal; May 2022
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title	LNCS
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference	IbPRIA
	Notes	MILAB; no menciona			Approved	no
	Call Number	Admin @ si @ BNR2022			Serial	3782
Permanent link to this record



	Author	Sergi Garcia Bordils; George Tom; Sangeeth Reddy; Minesh Mathew; Marçal Rusiñol; C.V. Jawahar; Dimosthenis Karatzas
	Title	Read While You Drive-Multilingual Text Tracking on the Road			Type	Conference Article
	Year	2022	Publication	15th IAPR International workshop on document analysis systems	Abbreviated Journal
	Volume	13237	Issue		Pages	756–770
	Keywords
	Abstract	Visual data obtained during driving scenarios usually contain large amounts of text that conveys semantic information necessary to analyse the urban environment and is integral to the traffic control plan. Yet, research on autonomous driving or driver assistance systems typically ignores this information. To advance research in this direction, we present RoadText-3K, a large driving video dataset with fully annotated text. RoadText-3K is three times bigger than its predecessor and contains data from varied geographical locations, unconstrained driving conditions and multiple languages and scripts. We offer a comprehensive analysis of tracking by detection and detection by tracking methods exploring the limits of state-of-the-art text detection. Finally, we propose a new end-to-end trainable tracking model that yields state-of-the-art results on this challenging dataset. Our experiments demonstrate the complexity and variability of RoadText-3K and establish a new, realistic benchmark for scene text tracking in the wild.
	Address	La Rochelle; France; May 2022
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title	LNCS
	Series Volume		Series Issue		Edition
	ISSN		ISBN	978-3-031-06554-5	Medium
	Area		Expedition		Conference	DAS
	Notes	DAG; 600.155; 611.022; 611.004			Approved	no
	Call Number	Admin @ si @ GTR2022			Serial	3783
Permanent link to this record



	Author	Francesco Pelosin; Saurav Jha; Andrea Torsello; Bogdan Raducanu; Joost Van de Weijer
	Title	Towards exemplar-free continual learning in vision transformers: an account of attention, functional and weight regularization			Type	Conference Article
	Year	2022	Publication	IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)	Abbreviated Journal
	Volume		Issue		Pages
	Keywords	Learning systems; Weight measurement; Image recognition; Surgery; Benchmark testing; Transformers; Stability analysis
	Abstract	In this paper, we investigate the continual learning of Vision Transformers (ViT) for the challenging exemplar-free scenario, with special focus on how to efficiently distill the knowledge of its crucial self-attention mechanism (SAM). Our work takes an initial step towards a surgical investigation of SAM for designing coherent continual learning methods in ViTs. We first carry out an evaluation of established continual learning regularization techniques. We then examine the effect of regularization when applied to two key enablers of SAM: (a) the contextualized embedding layers, for their ability to capture well-scaled representations with respect to the values, and (b) the prescaled attention maps, for carrying value-independent global contextual information. We depict the perks of each distilling strategy on two image recognition benchmarks (CIFAR100 and ImageNet-32) – while (a) leads to a better overall accuracy, (b) helps enhance the rigidity by maintaining competitive performances. Furthermore, we identify the limitation imposed by the symmetric nature of regularization losses. To alleviate this, we propose an asymmetric variant and apply it to the pooled output distillation (POD) loss adapted for ViTs. Our experiments confirm that introducing asymmetry to POD boosts its plasticity while retaining stability across (a) and (b). Moreover, we acknowledge low forgetting measures for all the compared methods, indicating that ViTs might be naturally inclined continual learners. 1
	Address	New Orleans; USA; June 2022
	Corporate Author				Thesis
	Publisher		Place of Publication		Editor
	Language		Summary Language		Original Title
	Series Editor		Series Title		Abbreviated Series Title
	Series Volume		Series Issue		Edition
	ISSN		ISBN		Medium
	Area		Expedition		Conference	CVPRW
	Notes	LAMP; 600.147			Approved	no
	Call Number	Admin @ si @ PJT2022			Serial	3784
Permanent link to this record