Skip to content

Latest commit

 

History

History
44 lines (28 loc) · 985 Bytes

README.md

File metadata and controls

44 lines (28 loc) · 985 Bytes
Raw

Team3

Week 1

Task 1: Compute image descriptors

We compute the histogram of 1000 bins of the images using different color spaces

  • RGB
  • XYZ
  • CIELAB
  • CIELUV
  • HSV
  • HLS
  • YCrCb
  • Grayscale

Task 2: Similarity measures

This are the similarity measures implemented in our code

  • Euclidean distance
  • L1 distance
  • X^2 distance
  • Histogram intersection
  • Hellinger kernel
  • Kullback-Leibler divergence
  • Jensen-Shannon divergence

Task 3: Compute similarities to museum images for images in QS1

We have tested for all different color spaces and similarity metrics.

Task 4: Evaluation QS1

For the query set 1 we have obtained an MAP@K factor of ...

Task 5: Remove back ground for QS2

To remove the background for images in query set 2 we have decided to look at the average value of the pixels at different depths from the boundaries of the global image. Those are the values we have used to segmentate our image.

Task 6: Evaluation QS2