1. Basic Flow:
Detect or sampel features
List of positions, scales, orientations
Associated list of dimensional descriptors
Index each one into pool of descriptors from previously seen images
Quantize to form bag of words fro the image.
2. Local feature corrrespondences:
3. Invariant local feature:
4. Corner detector
“A combined corner and edge detector”: C.Harris, M.Stephens, 1988.
5. Feature Matching
for each feature in one image, look at all the other features in the other images.
compute a short descriptor from each feature vector, or hash longer descriptors.
Nearest neighbour techniques
kd-trees and their variants.
Feature space outlier rejection:
Least squares fit
find average translation vector.
6. RANSAC algorithm for estimating
Random Sample Consensus
an algorithm for robust fiting of models in the presence of many data outliers
compare to robust statistics.