creating better thumbnails

29
Creating Better Thumbnails Chris Waclawik

Upload: ramiro

Post on 01-Feb-2016

35 views

Category:

Documents


0 download

DESCRIPTION

Creating Better Thumbnails. Chris Waclawik. Project Motivation. Thumbnails used to quickly select a specific a specific image from a set (when lacking appropriate metadata), but… Larger sets require smaller, less detailed thumbnails. Thumbnail effectiveness. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Creating Better  Thumbnails

Creating Better Thumbnails

Chris Waclawik

Page 2: Creating Better  Thumbnails

Project Motivation

• Thumbnails used to quickly select a specific a specific image from a set (when lacking appropriate metadata), but…

• Larger sets require smaller, less detailed thumbnails.

Page 3: Creating Better  Thumbnails

Thumbnail effectivenessCan you find the picture with the dog in it?

Page 4: Creating Better  Thumbnails

Thumbnail effectivenessCan you find the picture with the dog in it?

Page 5: Creating Better  Thumbnails

Thumbnail effectivenessCan you find the picture with the dog in it?

Page 6: Creating Better  Thumbnails

Problem Statement

• How can thumbnails be made more effective?

Page 7: Creating Better  Thumbnails

Show only the salient parts

Page 8: Creating Better  Thumbnails

Two subproblems

• How do we determine the salient portions of the image?

• Knowing the above, how should we crop the image?

Page 9: Creating Better  Thumbnails

Determining Image Saliency

• F. Liu, M. Gleicher, Region enhanced scale-invariant saliency detection, in: Proceedings of IEEE International Conference on Multimedia & Expo (ICME), 2006, pp. 1477-1480.

• Two-part method:– Scale-invariance makes this method suitable for

determining salient features of many sizes.– Region-enhancement aids in making logical crops

later on.

Page 10: Creating Better  Thumbnails

Scale-invariant Saliency

• Image transformed to perceptually uniform color space (Lu*v*).– Difference in color values proportional to perceived difference.

• Gaussian pyramid created and a contrast map is calculated for each level.– The contrast of each pixel is the sum of its distances from its

neighbors (in Lu*v* space).– Pixels closer to the center are given more weight.

• The contrast maps are summed to find scale invariant saliency.

• Algorithm implemented in Matlab.

Page 11: Creating Better  Thumbnails
Page 12: Creating Better  Thumbnails
Page 13: Creating Better  Thumbnails

Region-enhanced Saliency

• Regions extracted from the image using any preferred method.– Mean shift method used in paper (and in this

project).– C++ code from EDISON, Matlab wrapper function

from shawnlankton.com.• Scale-invariant saliency is averaged over each

region.

Page 14: Creating Better  Thumbnails
Page 15: Creating Better  Thumbnails

Saliency-based Cropping

• B. Suh, H. Ling, B. B. Bederson, and D. W. Jacobs, "Automatic Thumbnail Cropping and Its Effectiveness," in 16th Annual ACM Symposium on User Interface Software and Technology. 2003, pp. 95-104.

• What is the smallest crop that exceeds a specified saliency threshold (proportion of the total image saliency)? – Can be calculated by brute force, but inefficient

Page 16: Creating Better  Thumbnails

A greedy approach

• User specifies a target output size and saliency threshold.

• Select the most salient region.• Until we reach our threshold:

– Expand selection to include next most salient region.– Adjust selection to match aspect ratio of output size.

• Not guaranteed optimal crop– But much faster– Guaranteed to include whole regions

Page 17: Creating Better  Thumbnails

Demonstration

• Bounding box in red, correct aspect ratio in green

Page 18: Creating Better  Thumbnails
Page 19: Creating Better  Thumbnails

Determining Saliency Threshold Automatically

• Find the point of diminishing returns: when does increasing the crop area not become worth it?

• One simple method: select the image coming after the biggest increase in saliency.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1Saliency vs. Area

Saliency (percent of total)

Are

a (p

erce

nt o

f to

tal)

Page 20: Creating Better  Thumbnails

Results

Original Thumbnail

Page 21: Creating Better  Thumbnails

Results

Original Thumbnail

Page 22: Creating Better  Thumbnails

Results

Original Thumbnail

Page 23: Creating Better  Thumbnails

Results

Original Thumbnail

Page 24: Creating Better  Thumbnails

Results

Original Thumbnail

Page 25: Creating Better  Thumbnails

Results

Original Thumbnail

Page 26: Creating Better  Thumbnails

Results

Original Thumbnail

Page 27: Creating Better  Thumbnails

Results

Original Thumbnail

Page 28: Creating Better  Thumbnails

Results

Original Thumbnail

Page 29: Creating Better  Thumbnails

Further possible improvements

• Usability testing– Test users on how quickly they can locate a

specified image in a set given a table of thumbnails.

• Tweak region parameters• Saliency threshold selection

– Try out more sophisticated metrics• Basic feature recognition

– Face detection