Techniques for Image Classification, Object Detection and Object Segmentation

Reference:

Ville Viitaniemi and Jorma Laaksonen. Techniques for image classification, object detection and object segmentation. Technical Report TKK-ICS-R2, Helsinki University of Technology, Department of Information and Computer Science, Espoo, Finland, June 2008. Available online at http://www.cis.hut.fi/projects/cbir/.

Abstract:

In this paper we document the techniques which we used to participate in the PASCAL NoE VOC Challenge 2007 image analysis performance evaluation campaign. We took part in three of the image analysis competitions: image classication, object detection and object segmentation. In the classication task of the evaluation our method produced comparatively good performance, the 4th best of 19 submissions. In contrast, our detection results were quite modest. Our method's segmentation accuracy was the best of all submissions. Our approach for the classication task is based on fused classications by numerous global image features, including histograms of local features. The object detection combines similar classication of automatically extracted image segments and the previously obtained scene type classications. The object segmentations are obtained in a straightforward fashion from the detection results.

Keywords:

image classication, object detection, object segmentation, benchmarking

Suggested BibTeX entry:

@techreport{TKK-ICS-R2,
    address = {Espoo, Finland},
    author = {Ville Viitaniemi and Jorma Laaksonen},
    institution = {Helsinki University of Technology, Department of Information and Computer Science},
    month = {June},
    note = {Available online at \url{http://www.cis.hut.fi/projects/cbir/}.},
    number = {TKK-ICS-R2},
    pages = {18},
    title = {Techniques for Image Classification, Object Detection and Object Segmentation},
    type = {Technical Report},
    year = {2008},
}