Pacific Rim Conference on Multimedia

Advances in Multimedia Information Processing -- PCM 2015 pp 245-253

Convolutional Neural Networks Features: Principal Pyramidal Convolution

  • Yanming Guo
  • Songyang Lao
  • Yu Liu
  • Liang Bai
  • Shi Liu
  • Michael S. Lew
Conference paper

DOI: 10.1007/978-3-319-24075-6_24

Part of the Lecture Notes in Computer Science book series (LNCS, volume 9314)
Cite this paper as:
Guo Y., Lao S., Liu Y., Bai L., Liu S., Lew M.S. (2015) Convolutional Neural Networks Features: Principal Pyramidal Convolution. In: Ho YS., Sang J., Ro Y., Kim J., Wu F. (eds) Advances in Multimedia Information Processing -- PCM 2015. Lecture Notes in Computer Science, vol 9314. Springer, Cham

Abstract

The features extracted from convolutional neural networks (CNNs) are able to capture the discriminative part of an image and have shown superior performance in visual recognition. Furthermore, it has been verified that the CNN activations trained from large and diverse datasets can act as generic features and be transferred to other visual recognition tasks. In this paper, we aim to learn more from an image and present an effective method called Principal Pyramidal Convolution (PPC). The scheme first partitions the image into two levels, and extracts CNN activations for each sub-region along with the whole image, and then aggregates them together. The concatenated feature is later reduced to the standard dimension using Principal Component Analysis (PCA) algorithm, generating the refined CNN feature. When applied in image classification and retrieval tasks, the PPC feature consistently outperforms the conventional CNN feature, regardless of the network type where they derive from. Specifically, PPC achieves state-of-the-art result on the MIT Indoor67 dataset, utilizing the activations from Places-CNN.

Keywords

Convolutional neural networks Concatenate Principal component analysis Image classification Image retrieval 

Copyright information

© Springer International Publishing Switzerland 2015

Authors and Affiliations

  • Yanming Guo
    • 1
    • 2
  • Songyang Lao
    • 2
  • Yu Liu
    • 1
  • Liang Bai
    • 2
  • Shi Liu
    • 3
  • Michael S. Lew
    • 1
  1. 1.LIACS Media LabLeiden UniversityLeidenThe Netherlands
  2. 2.Science and Technology on Information Systems Engineering LaboratoryNational University of Defense TechnologyChangshaChina
  3. 3.School of Arts and MediaBeijing Normal UniversityBeijingChina

Personalised recommendations