2020/12/21 by Spyros Gidaris, Gidaris, Spyros, Andrei Bursuc +9 · 9 citations
Computer Science · #Advanced Image and Video Retrieval Techniques #Artificial intelligence #Bag-of-words model #Bag-of-words model in computer vision #Classifier (UML) #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Convolutional neural network #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Feature learning #Image (mathematics) #Image retrieval #Machine Learning (cs.LG) #Machine learning #Multimodal Machine Learning Applications #Natural language processing #Pascal (unit) #Pattern recognition (psychology) #Representation (politics) #Visual Word #Vocabulary #cs.CV #cs.LG
paper · pdf · doi:10.48550/arxiv.2012.11552
published in arXiv (Cornell University) (Cornell University) · Accepted to CVPR2021. Code at https://github.com/valeoai/obow
openalex publication_date 2020/12/21 · arxiv created 2021/10/29 · arxiv updated 2021/11/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/06
Learning image representations without human supervision is an important and active research field. Several recent approaches have successfully leveraged the idea of making such a representation invariant under different types of perturbations, especially via contrastive-based instance discrimination training. Although effective visual representations should indeed exhibit such invariances, there are other important characteristics, such as encoding contextual reasoning skills, for which alternative reconstruction-based approaches might be better suited. With this in mind, we propose a teacher-student scheme to learn representations by training a convolutional net to reconstruct a bag-of-visual-words (BoW) representation of an image, given as input a perturbed version of that same image. Our strategy performs an online training of both the teacher network (whose role is to generate the BoW targets) and the student network (whose role is to learn representations), along with an online update of the visual-words vocabulary (used for the BoW targets). This idea effectively enables fully online BoW-guided unsupervised learning. Extensive experiments demonstrate the interest of our BoW-based strategy which surpasses previous state-of-the-art methods (including contrastive-based ones) in several applications. For instance, in downstream tasks such Pascal object detection, Pascal classification and Places205 classification, our method improves over all prior unsupervised approaches, thus establishing new state-of-the-art results that are also significantly better even than those of supervised pre-training. We provide the implementation code at https://github.com/valeoai/obow.