Aim of the project

1
(No Transcript)
2
Aim of the project

• Take your image
• Submit it to the search engine
• Similar images are found
• Name of the building appears

3
Landmark Locations
4
Recognition of Buildings Using SIFT (ROBUST)
5
Solution overview

• Image representation by SIFT keypoints
• Sky and hedge detection
• Non object features are removed
• Similarity matching of features with database

6
Properties of SIFT features

• Scale invariance
• Rotation invariance
• Robustness to illumination changes
• Robustness to reasonable perspective
  transformation

7
SIFT Features
8
SIFT features computation

• Keypoint detection (local minima and maxima)
• Orientation of intensity gradients in the
  neighborhood
• Detection of significant orientations
• Histogram of orientations

9
Keypoint Refinement

• Sky detection (fast)
• Thresholding, morphological operations
• Applied to database and query images
• Hedge detection (slower)
• Hough transformation
• Closing morphology operation
• Applied to database images

10
Keypoints refinement
11
Similarity of features

• Matching between keypoints
• Take only reliable matches
• Similarity number of matched keypoints

12
Distance Matrix
13
Results

• SIFT Features outperformed PCA and Color
  Histogram Comparison
• Leave-one-out Validation
• First match correct 95.5
• First or second match correct 98.5

14
Demo Obscure Angle
15
Demo Rotated, Reflected
16
Demo People in Image
17
Demo Night
18
Demo Distant, Busy Scene
19
Demo Team