Image Classification with the Fisher Vector: Theory and Practice

Jorge Sanchez 1 Florent Perronnin 2 Thomas Mensink 3 Jakob Verbeek 4, *
* Corresponding author
4 LEAR - Learning and recognition in vision
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann, INPG - Institut National Polytechnique de Grenoble
Abstract : A standard approach to describe an image for classification and retrieval purposes is to extract a set of local patch descriptors, encode them into a high dimensional vector and pool them into an image-level signature. The most common patch encoding strategy consists in quantizing the local descriptors into a finite set of prototypical elements. This leads to the popular Bag-of-Visual words (BoV) representation. In this work, we propose to use the Fisher Kernel framework as an alternative patch encoding strategy: we describe patches by their deviation from an "universal" generative Gaussian mixture model. This representation, which we call Fisher Vector (FV) has many advantages: it is efficient to compute, it leads to excellent results even with efficient linear classifiers, and it can be compressed with a minimal loss of accuracy using product quantization. We report experimental results on five standard datasets - PASCAL VOC 2007, Caltech 256, SUN 397, ILSVRC 2010 and ImageNet10K - with up to 9M images and 10K classes, showing that the FV framework is a state-of-the-art patch encoding technique.
Document type :
Journal articles
Complete list of metadatas
Contributor : Jakob Verbeek <>
Submitted on : Wednesday, June 5, 2013 - 11:00:39 AM
Last modification on : Friday, August 2, 2019 - 3:24:01 PM
Long-term archiving on: Friday, September 6, 2013 - 4:10:57 AM


Files produced by the author(s)


  • HAL Id : hal-00830491, version 1


Jorge Sanchez, Florent Perronnin, Thomas Mensink, Jakob Verbeek. Image Classification with the Fisher Vector: Theory and Practice. International Journal of Computer Vision, Springer Verlag, 2013. ⟨hal-00830491v1⟩



Record views


Files downloads