1. Extract keypoints using Harris algorithm
C.G. Harris and M.J. Stephens. “A combined corner and edge detector“, Proceedings Fourth Alvey Vision Conference, Manchester. pp 147-151, 1988.
2. Extract keypoints using Laplacian of Gaussian (LoG) algorithm
Lindeberg, T. Feature Detection with Automatic Scale Selection, IEEE Transactions Pattern Analysis Machine Intelligence, 1998, 30, 77-116
3. Harris-Laplace algorithm
K. Mikolajczyk and C. Schmid. Scale & affine invariant interest point detectors. International Journal of Computer Vision, 2004
4. Gilles algorithm
S. Gilles, Robust Description and Matching of Images. PhD thesis, Oxford University, Ph.D. thesis, Oxford University, 1988.
The output keypoints is a matrix of dimension Nx2 or Nx3 with N the number of keypoints extracted.
– The first column gives the row poisition of the keypoints and
– The second column gives the column position of the keypoints.
– The third column gives the feature scale of the keypoints. This scale corresponds to the radius of the local neighborhood to consider.
- Resize the image to match the images in the database (640×480 pixels).
- Remove the background and clutter – HSV thresholding removes the large obstructing clutter. to isolate the object itself
- Enhance constrast: increasing dynamic range of image, amplify the small luminance changes within the origin image.
- Gamma correction amplifies the difference between very light and dark pixel values (blind amplification)
- The adaptive histogram equalisation – intelligently enhance the contrast of local region
- Contour influences the object that people see.
- Eight levels quantization – make contour more defined and flat.
- Feature detection.