Lect 18 - Object Recg ECE620 Summer 2011 Lec 2

Lect 18 - Object Recg ECE620 Summer 2011 Lec 2 - CVIP...

Info iconThis preview shows pages 1–13. Sign up to view the full content.

View Full Document Right Arrow Icon
1 CVIP Laboratory 1 Aly A. Farag University of Louisville Help with these slides were provided by Mostafa Abdelrahman Research Assistant, CVIP Lab, Electrical and Computer Engineering Dept., University of Louisville, KY, 40292 July 5 th , 2011 CVIP Laboratory Object Recognition & Feature Extraction Lecture 2
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
2 CVIP Laboratory 2 Outlines Object Recognition Problem o Definition o Application o Challenges Existing Object Recognition Techniques o Global features approach. o Local features approach. Edge detectors Corner Detectors Blob Detectors Region detectors SIFT CSIFT SPS
Background image of page 2
3 CVIP Laboratory 3 SIFT: Scale Invariant Feature Transform n David G. Lowe “Distinctive image features from scale -invariant keypoints ” (IJCV 2004)
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
4 CVIP Laboratory 4 Algorithm Get your interest points
Background image of page 4
5 CVIP Laboratory 5 In the absence of previous knowledge about the acquired objects, multi-scale representations appear as a solution for this ignorance. Several multi-scale representations have been developed in the literature. Quad-tree Pyramids Multi-Scale Representations
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
6 CVIP Laboratory 6 Scale-Space Representation For any image, two scale boundaries are defined to determine the scale range in which meaningful information can be obtained for a certain application: outer-scale and inner-scale. The outer-scale is defined as the minimum window size that completely contains the object of interest. The inner-scale is the scale at which the fine details of the object begin to appear. The inner scale can be interpreted as the image resolution .
Background image of page 6
7 CVIP Laboratory 7 Scale-Space Representation For computer vision and image processing applications, the proper scale range, the inner and the outer scales, cannot predefined. This is because of the lack of prior knowledge about the acquired images. Therefore, multi-scale representations are needed to cover several ranges of scales for the acquired image.
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
8 CVIP Laboratory 8 Multi-Scale Representations Pyramids Multi-scale representation using pyramids is preformed by successively smoothing and sub- sampling of the original signal (Burt’81 and Crowely’81). Coarser Scales Finer Scales
Background image of page 8