1 / 35

Fast multiresolution image querying

Fast multiresolution image querying. CS474/674 – Prof. Bebis. Case Study. C. Jacobs, A. Finkelstein, and D. Salesin, “Fast multiresolution image quering”, Proceedings of SIGGRAPH , pp. 277-286, 1995. Problem. Search an image database to retrieve images that are similar to a query image.

hildebrandt
Download Presentation

Fast multiresolution image querying

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Fast multiresolution image querying CS474/674 – Prof. Bebis

  2. Case Study • C. Jacobs, A. Finkelstein, and D. Salesin, “Fast multiresolution image quering”, Proceedings of SIGGRAPH, pp. 277-286, 1995

  3. Problem • Search an image database to retrieve images that are similar to a query image. “query by content” or “query by example” Typically, the K best matches are returned.

  4. Challenges • How to represent an image? • i.e., what features to use? • How to tolerate image distortions? • How to organize the images for fast retrieval? • How to reduce storage requirements?

  5. Image Distortions • This paper considers two types of image distortion: • A low-resolution image from a scanner or camera. • A rough sketch of the image painted by the user. painted low resolution target

  6. Traditional Metrics • Traditional metrics based on the L1and L2norms cannot handle inexact matching and are expensive to compute. Q: query T: target L1 L2 Experiments using these metrics have shown that the target image is in the highest 1% of the retrieved images only 3% of the time.

  7. Proposed Method: Key Ideas • Multi-resolution image decomposition using Haar wavelets. • Uses a metric that compares how many significant wavelet coefficients the query has in common with potential targets. • Efficient data organization to facilitate fast computation of the metric and speed-up search.

  8. User Interface Processes a 128 x 128 image query on a database of 20,000 images in under 0.5 seconds*. Returns 20 highest-ranked targets at interactive rates! *Faster processing times should be possible using current technology!

  9. Why using wavelets? • The use of wavelets allows the resolution of the query and target images to be different. • Wavelet decompositions are fast to compute and contain a small number of non-zero coefficients. • Wavelet-compressed images could be used directly.

  10. What color space to use? • Experimented with RGB, HSV, and YIQ color spaces. • Wavelet transform was applied on each color channel separately. • YIQ yielded the best performance.

  11. What wavelet to use? Haar wavelets are the fastest to compute and simplest to implement. Other types of wavelets might give better results but at a higher cost.

  12. What wavelet decomposition to use? • Experimented both with standard and non-standard decompositions for all three color spaces. • Standard decomposition worked best (i.e., both for scanned and painted queries). • Basis functions were normalized to become orthonormal to each other.

  13. First Idea: Coefficient Truncation • Keep only coefficients with large magnitude. • Improves discriminatory power of metric. • Improves speed and reduces storage requirements. • 128 x 128 image  1282 = 16,384 wavelet coefficients in each color channel. • The 60 largest coefficients in each channel worked best for painted queries (180 coefficients total). • The 40 largest coefficients in each channel worked best for scanned queries (120 coefficients total).

  14. Coefficient Truncation (cont’d) Truncated coefficients Wavelet decomposition White: positive coefficient Black: negative coefficient

  15. Second Idea: Coefficient Quantization • The mere presenceor absenceof large coefficients appears to be more important than their precise magnitude. • Improves speed and reduce storage requirements. • Improves discriminatory power of metric. • Quantize coefficients into three levels: +1, 0 and -1 • Large positive coefficients are quantized to +1 • Large negative coefficients are quantized to -1 • The rest are set to 0

  16. Components of the metric (cont’d) Truncated coefficients Truncated and quantized coefficients

  17. Third Idea: Wavelet-based Metric Q: Query and T: Target (single color channel) • Let Q[0, 0] and T[0, 0] be the scaling coefficients (i.e., average intensity of that channel). • Let and represent the truncated, quantized wavelet coefficients of Q and T (i.e., -1,0,1). Metric: wi,j : weights (determined statistically)

  18. Simplify the metric (1) Replace with (1 if true; 0 otherwise)

  19. Simplify the metric (cont’d) (2) Group terms together into "buckets" so that only a smaller number of weights wi,jis needed. i,j

  20. Simplify the metric (cont’d) • The function bin(i, j) groups different coefficients into a small number of bins (i.e., 6 bins per color channel): bin(i, j) = min(max(i, j), 5) • Each bin is weighted by w[b] • Weights were determined using a statistical test. 0 1 2 3 4 5 0 1 2 3 4 5

  21. Simplify the metric (cont’d) (3) Consider only the terms for which • Leads to even faster computations. • Allows for a query without much detail to match a detailed target image. i,j

  22. Simplify the metric (cont’d) (4) Count the number of matching coefficients than the number of mismatching coefficients. • The majority of database images will not match the query. • Use “search arrays” to implement this idea efficiently. (1 if true; 0 otherwise)

  23. Simplify the metric (cont’d)

  24. Simplify the metric (cont’d) • The term does not depend on the target image and can be ignored: Final Metric

  25. Example

  26. Search Arrays • Use a set of six 2D arrays (i.e. search arrays) to organize the mcoefficients from every target image T. • There is an array for every combination of sign (+ or -) and color channel (Y, I, and Q): contains a list of all target images T having a large positive wavelet coefficient at [i, j] location, in color channel c.

  27. Search Arrays (cont’d) T1, T9, T22, … T3, T9, T40, … … Compute a score for each target image found in T1, T22, T98, … …

  28. Querying Using Search Arrays • Steps (for each color channel c) • Compute the difference between the query’s average intensity Qc[0, 0] and those in the database. (2) For each of the m nonzero, truncated wavelet coefficients Qc[i, j], go through the list corresponding to Dc+[i, j] or Dc- [i, j] (i.e., depending on the sign of Qc[i, j]). (3) Update the score of each target image found in those lists.

  29. Algorithm • Preprocessing • Perform a standard 2D Haar wavelet decomposition of every image in the database. (2) Store T[0,0] for each color channel as well as the indices and signs of the mlargestwavelet coefficients (2D search arrays).

  30. Algorithm (cont’d) • Querying • Perform a standard 2D Haar wavelet decomposition on the query image. (2) Compute T[0,0] for each color channel as well as the indices and signs of the mlargestwavelet coefficients (3) Compute the score of each target image using: (4) Return 20 highest-ranked target images

  31. Examples • Query examples using painted/scanned queries (ranks for database sizes: 1093 | 20,558)

  32. Examples (cont’d) Interactive query using painted query: (ranks for database sizes: 1093 | 20,558)

  33. Success Rate Lq : proposed metric Percentage of queries whose correct target was ranked among the top 1% of images in a database of 1093 images.

  34. Time requirements Lq : proposed metric Average times to match a single query in a database of 1093/20,558 images.

  35. Extension • V. Nikulin and G. Bebis, "Multiresolution Image Retrieval Through Fusion", SPIE Electronic Imaging (Storage and Retrieval Methods and Applications for Multimedia), San Jose, January 2004.

More Related