Seam Carving

Content-aware image-resizing

Transcript

1. Seam Carving Edwin Kwok Thursday, 30 August, 12
2. In One Line... • Method for Content-Aware Image Resizing Thursday, 30 August, 12
3. Thursday, 30 August, 12
4. Different Displays-Different Resolutions-(Different Images??) Thursday, 30 August, 12
5. HTML allows automatic layout of the text, but not the images Thursday, 30 August, 12
6. Solutions • Crop • Scale • Segmentation Thursday, 30 August, 12
7. But we want... • Content is more important • don’t resize the content please Thursday, 30 August, 12
8. Math Thursday, 30 August, 12
9. Definition • Goal is to remove unnoticeable pixels. • Energy Function: • Remove Pixels with least energy • Sort pixels by Energy, remove smallest pixels (optimum) • Remove Column with smallest total Energy (column) • Remove Seam with smallest total Energy (seam) Thursday, 30 August, 12
10. Seam • Vertical Seam: • an 8-connected path of pixels in the image from top to bottom, containing one, and only one, pixel in each row of the image Thursday, 30 August, 12
11. Seam • Energy of a Seam • Minimum Energy Seam Thursday, 30 August, 12
12. Pixel Removal Criteria • Optimal Thursday, 30 August, 12
13. Pixel Removal Criteria • Column Thursday, 30 August, 12
14. Pixel Removal Criteria • Column Thursday, 30 August, 12
15. Pixel Removal Criteria • Comparison Thursday, 30 August, 12
16. Aspect Ratio Change • Change the image from (n x m) to (n x m’) • (assume m’ < m) • Remove m-m’=c seams successively Seam Carving Scaling Thursday, 30 August, 12
17. Optimal Seam Ordering • Change the image from (n x m) to (n’ x m’) • (assume m’ < m and n’ < n) • What to do? What is the correct order of seam carving? • Remove vertical seams first? • Horizontal seams first? • Or alternate between the two? Thursday, 30 August, 12
18. Optimal Seam Ordering • We optimize • where k=r+c, r=(m-m’), c=(n-n’) and αi is used as a parameter that determine if at step i we remove a horizontal or vertical seam: α={0,1} Thursday, 30 August, 12
19. Algorithm • Dynamic programming Thursday, 30 August, 12
20. Application Thursday, 30 August, 12
21. Image Enlarging • compute the optimal vertical (horizontal) seam ‘s’ on I • duplicate the pixels of s by averaging them with their left and right neighbors (top and bottom in the horizontal case) Thursday, 30 August, 12
22. Image Enlarging • To reduce the width by ‘k’, compute first ‘k’ vertical seams and duplicate them • All seams duplicated, 50% enlargement. For more do it in steps. Thursday, 30 August, 12
23. Content Amplification • Scale the image, this will scale everything, content as well as non-content. • Shrink the scaled-image by seam carving, this will carve out the non-content part. Thursday, 30 August, 12
24. Object Removal • User marks the target object to be removed. • Then seams are removed from the image until all marked pixels are gone. • To regain the original size, seam insertion is employed. Thursday, 30 August, 12
25. Photo Hunt... Thursday, 30 August, 12
26. Photo Hunt... Thursday, 30 August, 12
27. Object Removal • Object marking to remove unwanted results Thursday, 30 August, 12
28. Limitation • Methods mentioned so far are not real-time. • We calculate each seam remove it and calculate the next seam. • Too much content • No space for seam to avoid content Thursday, 30 August, 12
29. Link • Paper • ftp://ftp1.idc.ac.il/Arik_shamir/SCweb/imret/vidret.pdf • Dynamic Programming for Seam Carving Algorithm • http://w3.impa.br/~rdcastan/SeamWeb/Seam%20Carving%20Part %201.pdf • Youtube demo • http://www.youtube.com/watch?v=6NcIJXTlugc • Search google for Seam Carving…. Many many blogs… Thursday, 30 August, 12
30. Thank you Thursday, 30 August, 12