Training Vision Transformers for Image Retrieval - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Preprints, Working Papers, ... Year : 2022

Training Vision Transformers for Image Retrieval


Transformers have shown outstanding results for natural language understanding and, more recently, for image classification. We here extend this work and propose a transformer-based approach for image retrieval: we adopt vision transformers for generating image descriptors and train the resulting model with a metric learning objective, which combines a contrastive loss with a differential entropy regularizer. Our results show consistent and significant improvements of transformers over convolutionbased approaches. In particular, our method outperforms the state of the art on several public benchmarks for category-level retrieval, namely Stanford Online Product, In-Shop and CUB-200. Furthermore, our experiments on ROxford and RParis also show that, in comparable settings, transformers are competitive for particular object retrieval, especially in the regime of short vector representations and low-resolution images.
Fichier principal
Vignette du fichier
irt_haltools.pdf (3.37 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03572734 , version 1 (14-02-2022)


  • HAL Id : hal-03572734 , version 1


Alaaeldin El-Nouby, Natalia Neverova, Ivan Laptev, Hervé Jégou. Training Vision Transformers for Image Retrieval. 2022. ⟨hal-03572734⟩
121 View
249 Download


Gmail Facebook X LinkedIn More