Centre de documentation
scientifique

A shape transformation-based dataset augmentation framework for pedestrian detection / Zhe Chen in International journal of computer vision, vol 129 n° 4 (April 2021)  

Public [article]  inInternational journal of computer vision > vol 129 n° 4 (April 2021) . - pp 1121 - 1138 Titre : A shape transformation-based dataset augmentation framework for pedestrian detection Type de document : Article/Communication Auteurs : Zhe Chen, Auteur ; Wanli Ouyang, Auteur ; Tongliang Liu, Auteur ; et al., Auteur Année de publication : 2021 Article en page(s) : pp 1121 - 1138 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] apprentissage par transformation [Termes IGN] apprentissage profond [Termes IGN] déformation d'objet [Termes IGN] détection de piéton [Termes IGN] jeu de données [Termes IGN] synthèse d'image [Termes IGN] vision par ordinateur Résumé : (auteur) Deep learning-based computer vision is usually data-hungry. Many researchers attempt to augment datasets with synthesized data to improve model robustness. However, the augmentation of popular pedestrian datasets, such as Caltech and Citypersons, can be extremely challenging because real pedestrians are commonly in low quality. Due to the factors like occlusions, blurs, and low-resolution, it is significantly difficult for existing augmentation approaches, which generally synthesize data using 3D engines or generative adversarial networks (GANs), to generate realistic-looking pedestrians. Alternatively, to access much more natural-looking pedestrians, we propose to augment pedestrian detection datasets by transforming real pedestrians from the same dataset into different shapes. Accordingly, we propose the Shape Transformation-based Dataset Augmentation (STDA) framework. The proposed framework is composed of two subsequent modules, i.e. the shape-guided deformation and the environment adaptation. In the first module, we introduce a shape-guided warping field to help deform the shape of a real pedestrian into a different shape. Then, in the second stage, we propose an environment-aware blending map to better adapt the deformed pedestrians into surrounding environments, obtaining more realistic-llooking pedestrians and more beneficial augmentation results for pedestrian detection. Extensive empirical studies on different pedestrian detection benchmarks show that the proposed STDA framework consistently produces much better augmentation results than other pedestrian synthesis approaches using low-quality pedestrians. By augmenting the original datasets, our proposed framework also improves the baseline pedestrian detector by up to 38% on the evaluated benchmarks, achieving state-of-the-art performance. Numéro de notice : A2021-354 Affiliation des auteurs : non IGN Thématique : IMAGERIE/INFORMATIQUE Nature : Article DOI : 10.1007/s11263-020-01412-0 Date de publication en ligne : 09/01/2021 En ligne : https://doi.org/10.1007/s11263-020-01412-0 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97606 [article]