Presentation is loading. Please wait.

Presentation is loading. Please wait.

資訊碩一 10077034 蔡勇儀 Date : 603.  Introduction  Basic Theory  Application & Implementation ◦ Aspect Ratio Change ◦ Retargeting with Optimal.

Similar presentations


Presentation on theme: "資訊碩一 10077034 蔡勇儀 Date : 603.  Introduction  Basic Theory  Application & Implementation ◦ Aspect Ratio Change ◦ Retargeting with Optimal."— Presentation transcript:

1 資訊碩一 10077034 蔡勇儀 Date : 2012/01/03 @LAB 603

2  Introduction  Basic Theory  Application & Implementation ◦ Aspect Ratio Change ◦ Retargeting with Optimal Seams-Order ◦ Enlarging ◦ Content Amplification ◦ Seam Carving in the gradient domain ◦ Object Removal  Multi-size Images  Limitation  Conclusions and Future Work

3  Muti-Media and Embedding System(e.g. Cell Phone) grow fast , Resize or Multi-Size scaling are more important than past.  Standard image scaling is not sufficient since it is oblivious to the image content and typically can be applied only uniformly.  For improve the problem, many researcher prove some good idea.

4  The following is main methods for scaling : ◦ Corp (figure2(b)) ◦ Column or Row removal (figure2(c)) ◦ Pixel energy removal (figure2(e)) ◦ Optimal Pixel energy removal (figure2(f)) ◦ Object detection ◦ Seam Craving (figure2(d))  We can found the seam method have better result!

5  Introduction  Basic Theory  Application & Implementation ◦ Aspect Ratio Change ◦ Retargeting with Optimal Seams-Order ◦ Enlarging ◦ Content Amplification ◦ Seam Carving in the gradient domain ◦ Object Removal  Multi-size Images  Limitation  Conclusions and Future Work

6  Step1 – Find seam ◦ Find a path which have the minimum energy sum from image top to bottom.  Step2 – Remove the Min. seam ◦ When found all seam, select the Min. seam remove.  Step3 – Repeat above step until get the demand size

7  Give an energy function  Define Seam  Define the Seam Cost

8  Find the minimum seam  Remove S* form image and lnstead of neighbors

9  Repeat above step until get the demand size

10  What’s energy function is the best? ◦ e1 ◦ Entropy  9*9 Windows add to e1 ◦ Segmentation ( add to e1) ◦ Histogram of Gradients  11*11cell around a pixel, 8-bins

11  They all accommodate a similar range for resizing.  We found either e1 or eHoG to work quite well.

12  Introduction  Basic Theory  Application & Implementation ◦ Aspect Ratio Change ◦ Retargeting with Optimal Seams-Order ◦ Enlarging ◦ Content Amplification ◦ Seam Carving in the gradient domain ◦ Object Removal  Multi-size Images  Limitation  Conclusions and Future Work

13  Only one axis adjust  A picture size n*m  n*m’ or n’*m  n >= n’ m >= m’  Remove n-n’ or m-m’ seams  Enlarge at other page

14  What’s the optimal order for remove seams? Column or Row or Other? How could decide?  Using dynamic programming ◦ where k = r+c, c = (m−m’), r = (n−n’) ◦ α i is used as a parameter  that determine if at step i we remove a horizontal or vertical seam:

15  Define transport map T ◦ T(r,c)=min(T(r-1,c)+E(s y (I n-r+1*m-c )), T(r,c-1)+E(s x (I n-r*m-c+1 )) ) ◦ where I n-r*m-c  denotes an image of size (n−r)×(m−c), ◦ E(s x (I)) and E(s y (I))  are the cost of the respective seam removal operation.  Build the 1 bit map for record the direction

16

17  When m’ > m or n’ > n, we should insert seams to the picture.  Find the smallest energy seam for copy and insert, repeat until equal the demand scale.  But…

18  Every time found the same seam, so we should decide all seams which need copy at first.  If m’ > m then we need insert (m’-m) seams. Find them and copy it for insertation.

19 The origin picture ScalarSeam

20  Using same scalar enlarge then use seams- carving for recover to the origin size.

21  If energy funciton use the gradient, then color show at remove place will be more nature after seam carving.

22  User mark the part which want to remove.  Decrease the energy on the part is removed.  Insert seams for keeping origin size.

23  Introduction  Basic Theory  Application & Implementation ◦ Aspect Ratio Change ◦ Retargeting with Optimal Seams-Order ◦ Enlarging ◦ Content Amplification ◦ Seam Carving in the gradient domain ◦ Object Removal  Multi-size Images  Limitation  Conclusions and Future Work

24  User want find the optimal picture scalar for their demand, so we need the real time opreation.  But the picture’s size 400*500 to 100*100 in about 2.2 seconds, it is too long to real time. How could do for real time?

25  Make the index map for seams before user operation.  Build the horizontal & vertical index map (H&V)  But there will a big problem for operation that is H & V will be collided.  The sample solution is decide one just do one direction and then other direction need degenerate the index and redo the select seams operation

26

27  Introduction  Basic Theory  Application & Implementation ◦ Aspect Ratio Change ◦ Retargeting with Optimal Seams-Order ◦ Enlarging ◦ Content Amplification ◦ Seam Carving in the gradient domain ◦ Object Removal  Multi-size Images  Limitation  Conclusions and Future Work

28  this method ◦ does not work automatically on all images. ◦ can be corrected by adding higher level cues, either manual or automatic. Figure 14, Figure 15  Other times, ◦ not even high level information can solve the problem.  two major factors that limit this seam carving approach. ◦ The first  is the amount of content in an image.  If the image is too condensed,  it does not contain ‘less important’ areas,  then any type of content-aware resizing strategy will not succeed. ◦ The second type of limitation  is the layout of the image content.  In certain types of images, albeit not being condensed,the content is laid out in a manner that prevents the seams to bypass important

29

30  Introduction  Basic Theory  Application & Implementation ◦ Aspect Ratio Change ◦ Retargeting with Optimal Seams-Order ◦ Enlarging ◦ Content Amplification ◦ Seam Carving in the gradient domain ◦ Object Removal  Multi-size Images  Limitation  Conclusions and Future Work

31  to extend this approach to other domains, ◦ the first of which would be resizing of video. ◦ Since there are cases  when scaling can achieve better results for resizing,  would like to investigate the possibility  to combine the two approaches,  Specifically to define more robust multi-size images. ◦ would also like to find a better way  to combine horizontal and vertical seams in multi-size images.

32  Q&A


Download ppt "資訊碩一 10077034 蔡勇儀 Date : 603.  Introduction  Basic Theory  Application & Implementation ◦ Aspect Ratio Change ◦ Retargeting with Optimal."

Similar presentations


Ads by Google