Aaju baaju frames might come in handy for noise reduction... averaging and all... esp I frames... cant see any relation with region or edge growing but had totally 4gotten we were working on a video!
Prateek
Wednesday, November 25, 2009
Saturday, November 21, 2009
disaster
Bad bad cluster mean values found by the region growing algo... That's when we neglected edge pixels as they usually have gradient in the transition between colours...
Leaving this for now (Dunno why I started today, already have enough things to screw up till Monday :P)...
Leaving this for now (Dunno why I started today, already have enough things to screw up till Monday :P)...
Thursday, November 19, 2009
trying to get better ede detection
While earlier we had tested
image -> sobel filter output -> grayscaling -> thresholding (-> median filter output) -> region partitioning,
on reordering the transforms applied as:
image -> median filter output -> sobel filter output -> grayscaling -> thresholding -> region partitioning
we find that
1. the number of broken edges is reduced resulting in better partitioning.
2. However, detail present in the thresheld and partitioned outputs is significantly low as well.
3. The second method works well even for very low values of threshold (23 was used instead of 63).
Detail may be added in preprocessing or the result, using detail obtained by wavelet transform. This is yet to be tested... for now it is time to go visit our guide and ask him for suggestions for localized spatial clustering, things like kernel methods in k-means might actually be relevant and will be assessed.
Prateek
image -> sobel filter output -> grayscaling -> thresholding (-> median filter output) -> region partitioning,
on reordering the transforms applied as:
image -> median filter output -> sobel filter output -> grayscaling -> thresholding -> region partitioning
we find that
1. the number of broken edges is reduced resulting in better partitioning.
2. However, detail present in the thresheld and partitioned outputs is significantly low as well.
3. The second method works well even for very low values of threshold (23 was used instead of 63).
Detail may be added in preprocessing or the result, using detail obtained by wavelet transform. This is yet to be tested... for now it is time to go visit our guide and ask him for suggestions for localized spatial clustering, things like kernel methods in k-means might actually be relevant and will be assessed.
Prateek
Saturday, November 7, 2009
reading frame data into ppm
using http://web.me.com/dhoerl/Home/Tech_Blog/Entries/2009/1/22_Revised_avcodec_sample.c.html and some help from pv and http://facweb.iitkgp.ernet.in/~pds/notes/array.html#arrayandpointer got through to reading frame data and understanding how the AVFrame->data is to be traversed for pixel data.
for each frame
for each frame
for(y=0; y<height; y++) { for (zz=0;zz<width;++zz) { red = *(pFrame->data[0]+y*pFrame->linesize[0]+zz*3+0)); green = *(pFrame->data[0]+y*pFrame->linesize[0]+zz*3+1)); blue = *(pFrame->data[0]+y*pFrame->linesize[0]+zz*3+2)); alpha = *(pFrame->data[0]+y*pFrame->linesize[0]+zz*3+2)); } }
Wednesday, October 21, 2009
Re: video capture library
yea we'd used openCV with ffmpeg support i remember :P... we can use the same na? even hav the code at the tracker site
Prateek
Prateek
On Wed, Oct 21, 2009 at 8:41 PM, Sayandeep Purkayasth <deepcyan@gmail.com> wrote:
we used some other library. anyway, i was looking at the second tutorial and it seems usable... much of the code is already written.On Wed, Oct 21, 2009 at 8:37 PM, Hollow Man <evanescentpv@gmail.com> wrote:
by capture if u mean storing every frame as an image, havent we done that already (7th sem) using ffmpeg?
Prateek
On Wed, Oct 21, 2009 at 8:23 PM, Sayandeep Purkayasth <deepcyan@gmail.com> wrote:what say we use ffmpeg this time?
if so, find appropriate tutorials http://www.inb.uni-luebeck.de/~boehme/using_libavcodec.html (outdated code) and http://www.dranger.com/ffmpeg/tutorial01.html (not checked yet)
On Wed, Oct 21, 2009 at 4:44 PM, Sayandeep Purkayasth <deepcyan@gmail.com> wrote:one promising one
- Simple, Thread-safe Approximate Nearest Neighbor (STANN) C++
Re: video capture library
we used some other library. anyway, i was looking at the second tutorial and it seems usable... much of the code is already written.
On Wed, Oct 21, 2009 at 8:37 PM, Hollow Man <evanescentpv@gmail.com> wrote:
by capture if u mean storing every frame as an image, havent we done that already (7th sem) using ffmpeg?
Prateek
On Wed, Oct 21, 2009 at 8:23 PM, Sayandeep Purkayasth <deepcyan@gmail.com> wrote:what say we use ffmpeg this time?
if so, find appropriate tutorials http://www.inb.uni-luebeck.de/~boehme/using_libavcodec.html (outdated code) and http://www.dranger.com/ffmpeg/tutorial01.html (not checked yet)
On Wed, Oct 21, 2009 at 4:44 PM, Sayandeep Purkayasth <deepcyan@gmail.com> wrote:one promising one
- Simple, Thread-safe Approximate Nearest Neighbor (STANN) C++
Re: video capture library
by capture if u mean storing every frame as an image, havent we done that already (7th sem) using ffmpeg?
Prateek
Prateek
On Wed, Oct 21, 2009 at 8:23 PM, Sayandeep Purkayasth <deepcyan@gmail.com> wrote:
what say we use ffmpeg this time?
if so, find appropriate tutorials http://www.inb.uni-luebeck.de/~boehme/using_libavcodec.html (outdated code) and http://www.dranger.com/ffmpeg/tutorial01.html (not checked yet)
On Wed, Oct 21, 2009 at 4:44 PM, Sayandeep Purkayasth <deepcyan@gmail.com> wrote:one promising one
- Simple, Thread-safe Approximate Nearest Neighbor (STANN) C++
Subscribe to:
Posts (Atom)