Unsupervised Pre-Training of Image Features on Non-Curated Data - Archive ouverte HAL Access content directly
Conference Papers Year : 2019

Unsupervised Pre-Training of Image Features on Non-Curated Data


Pre-training general-purpose visual features with convolutional neural networks without relying on annotations is a challenging and important task. Most recent efforts in unsupervised feature learning have focused on either small or highly curated datasets like ImageNet, whereas using non-curated raw datasets was found to decrease the feature quality when evaluated on a transfer task. Our goal is to bridge the performance gap between unsupervised methods trained on curated data, which are costly to obtain, and massive raw datasets that are easily available. To that effect, we propose a new unsupervised approach which leverages self-supervision and clustering to capture complementary statistics from large-scale data. We validate our approach on 96 million images from YFCC100M [42], achieving state-of-the-art results among unsupervised methods on standard benchmarks, which confirms the potential of unsupervised learning when only non-curated raw data are available. We also show that pre-training a supervised VGG-16 with our method achieves 74.9% top-1 classification accuracy on the validation set of ImageNet, which is an improvement of +0.8% over the same network trained from scratch. Our code is available at https://github.com/facebookresearch/DeeperCluster.
Fichier principal
Vignette du fichier
main.pdf (2.22 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-02119564 , version 1 (03-05-2019)
hal-02119564 , version 2 (09-09-2019)



Mathilde Caron, Piotr Bojanowski, Julien Mairal, Armand Joulin. Unsupervised Pre-Training of Image Features on Non-Curated Data. ICCV 2019 - International Conference on Computer Vision, Oct 2019, Seoul, South Korea. pp.2959-2968, ⟨10.1109/ICCV.2019.00305⟩. ⟨hal-02119564v2⟩
263 View
939 Download



Gmail Mastodon Facebook X LinkedIn More