1. Project Bid
I've been invited to bid on a project... here's the basics:
A program that will allow a home designer to see the "after" before the
work is done.
- User will take digital picture of client's location.
- User will apply patterns to select portions of digital picture to
represent what it will look like after the work is done.
Seems like a simple enough app (famous last words, I know)... What do
y'all think about it? Who wants to help (paying job)? I could also serve
as go-between if you'd like to bid on the project.
2. Re: Project Bid
posted by: (not specified)
Sounds simple, but it isn't.
Home depot, and some of the bigger paint manufacturers, have done
this already, perhaps you should check them out before re-inventing
the wheel.
Irv
3. Re: Project Bid
Guest wrote:
>Sounds simple, but it isn't.
>
>
I've thought about it and didn't think it would be easy, but as I
thought more, I thought it could be pretty simple... Take the picture
with three "markers" which can be used to determine the slope of the
plane and area covered, tile the pattern onto a large block, do 3D
rotate according to slope specs, then apply over specified area of
photograph... Apply lighting as determined from the original pic. BAM!
"After Preview."
>Home depot, and some of the bigger paint manufacturers, have done
>this already, perhaps you should check them out before re-inventing
>the wheel.
>
>
The bid spec wants a brandable software product for distribution to
contractors... I don't think I'll be able to do anything with Home
Depot's software. :/ Unless you know of something out there that does
this already! :)
4. Re: Project Bid
C. K. Lester wrote:
> I've been invited to bid on a project... here's the basics:
I've got two questions about it:
1. How does the use select an area from the photo? Do they set points that
connect into a solid polygon (trivial) or expect to be able to click some
feature (like one of the walls) and have the program automatically determine
the borders (nontrivial).
2. How much 3D information is supposed to be gleaned from the image? For
example, is the pattern applied as a flat floodfill (trivial), scalable
floodfill where the user selected how magnified the pattern is (trivial), or
is the program supposed to apply the pattern with a 3D effect where (for
example) a pattern becomes larger as it comes "closer" (nontrivial).
-- David Cuny
5. Re: Project Bid
posted by: euphoric at cklester.com
I don't know how to "reply" to a message on the web interface... Can somebody
give me a clue? :)
David C. asked, "How does the use select an area from the photo? Do they set
points that connect into a solid polygon (trivial)"
Yes. I imagine them either "painting" the area (creating a solid mask), drawing
around it, whatever. There might could be a "wand" that can try and determine
what the area is, but that's more complicated than I want to get.
David C. asked, "How much 3D information is supposed to be gleaned from the
image?"
I propose the designer sets three "markers" (in a triangle pattern, or one
triangular marker with three points) in the area and measures the distance from
the camera to each, as well as the height off the floor of the camera. Using that
data, the slope of the area can be determined. The computer would then be able to
simply apply a texture to a square, form the square to be parallel to the plane,
and use the mask to apply the 3D tile on the surface. With Win32DIB, this sounds
very easily doable.
Right? :)
6. Re: Project Bid
posted by: Tommy Carlier < tommy.carlier at pandora.be >
If you use Win32Dib, you can do it like this:
- user selects a wall with a polygon-selection tool (limited to 4 points).
- user selects a pattern.
- application first makes a new bitmap with a width and height of the rectangle
surrounding the polygon (the polygon has to be totally inside the rectangle).
- application paints the pattern onto that bitmap (tileDibToDib).
- application paints this bitmap as a transformed polygon (the selected polygon)
onto the original picture (drawDibTransformedPolygon).
That way, the pattern has the same perspective as the selected wall. In the next
version of Win32Dib, I'll probably add alpha-blending of bitmaps, which means you
can blend the original lighting of the wall with the new pattern.
7. Re: Project Bid
TC wrote:
>If you use Win32Dib, you can do it like this:
>
>- user selects a wall with a polygon-selection tool (limited to 4 points).
>
>
I'll probably have the user "paint" the area desired. That way, it can
get real specific. This is going to be needed for floors, also, where
all sorts of furniture items, potted plants, etc., might be found and
for which a 4-point polygon would not work.
>- user selects a pattern.
>- application first makes a new bitmap with a width and height of the rectangle
>surrounding the polygon (the polygon has to be totally inside the rectangle).
>- application paints the pattern onto that bitmap (tileDibToDib).
>- application paints this bitmap as a transformed polygon (the selected
>polygon) onto the original picture (drawDibTransformedPolygon).
>
>
Yeah, that's the way I envisioned it working. No problem! (The last step
would of course use the mask created in the first step.)
>That way, the pattern has the same perspective as the selected wall. In the
>next version of Win32Dib, I'll probably add alpha-blending of bitmaps, which
>means you can blend the original lighting of the wall with the new pattern.
>
>
THAT I can use... bring it on!!! :)
8. Re: Project Bid
I just uploaded a new version of Win32Dib (0.4.0) to my website
(http://users.pandora.be/tommycarlier/eu). This new version has some features
that might be useful for your project:
-> translucency: drawDibToDib and drawTransDibToDib (transparent bitmap) now
have an extra argument 'alpha'. This is an integer between 0 and 255. 0 means
fully transparent, 255 is fully opaque, and anything in between is translucent
(partially transparent). Check out the demo tiledib.exw to see the effect.
-> applyDibBrightnessToDib: this new procedure is just like drawDibToDib, but it
doesn't draw the source bitmap colors to the destination bitmap: it just copies
its brightness/lighting. You could use this to apply a pattern to a flat-shaded
bitmap. In your project, you could apply the original lighting of the wall to the
new pattern. Check out the demo manip.exw: 'Apply Brightness' applies the
brightness of a bitmap (containing horizontal lines) to the original bitmap.
PS: EuMario: I haven't added a bit depth reducer (yet), but now there is a
saveDibGray-function that saves a bitmap as an 8-bit grayscale BMP-file. The
advantage: smaller file + you don't have to make the bitmap gray yourself (bitmap
itself stays intact).
