|
Well, the ShearX and ShearY values would be here:
1, Y, 0,
X, 1, 0,
0, 0, 1
If you already know how to apply shear to a matrix, whats so hard about getting the values? Am i completely missing something?
Infact, how are you getting the other values? If i were to scale along the x-axis by 2, and the y-axis by 3. Then translate by (5,3) i would end up with this matrix:
2, 0, 0
0, 3, 0
10,9, 1
But if i translate then scale we just get:
2, 0, 0
0, 3, 0
5, 3, 1
So how can you know what order transformations where done in, without knowing what transformations where applied and vice versa?
My current favourite word is: I'm starting to run out of fav. words!
-SK Genius
Game Programming articles start - here[ ^]-
|
|
|
|
|
SK Genius wrote: Infact, how are you getting the other values?
Good question!
Here's the code for extracting the values. Note I got this from another lib, so I'm not 100% clear on all the math - I *think* I understand the idea behind it though.
double matrix_[3][3];
double Matrix2D::getRotation() const
{
double x1 = 0.0;
double y1 = 0.0;
double x2 = 1.0;
double y2 = 0.0;
apply(x1, y1);
apply(x2, y2);
return atan2(y2-y1, x2-x1);
}
void Matrix2D::getScale( double& x, double& y ) const
{
double x1 = 0.0;
double y1 = 0.0;
double x2 = 1.0;
double y2 = 1.0;
Matrix2D rot;
rot.rotate( -getRotation() );
rot.multiply( *this );
rot.apply(x1, y1);
rot.apply(x2, y2);
x = x2 - x1;
y = y2 - y1;
}
note that apply() simply transforms the (x,y) values by whatever is in the matrix_ array.
¡El diablo está en mis pantalones! ¡Mire, mire!
Real Mentats use only 100% pure, unfooled around with Sapho Juice(tm)!
SELECT * FROM User WHERE Clue > 0
0 rows returned
Save an Orange - Use the VCF!
VCF Blog
|
|
|
|
|
I see, so you can get the rotation without worrying about scale or translate etc, because the rotation is completely separate. To get the scale you have to undo the rotation, so you can pretty much just get the values of X and Y here:
X, 0, 0
0, Y, 0
0, 0, 1
Shear and translate are in different parts of the matrix.
The problem is, both shear and translate are affected by Scale, so you'd have to know if the matrix was scaled before or after a shear or translate was applied (if at all).
Rotation isn't affected by scale because the angle remains the same, and the points just get further from the origin. So, the best you can do is to undo the rotation like you have in getScale() and just get the values from the matrix, even thought they may have been scaled.
So, using the apply method (rather than just using the array, i don't know if you have direct access to it):
void Matrix2D::getShear( double& x, double& y ) const
{
double x1 = 1.0;
double y1 = 1.0;
Matrix2D rot;
rot.rotate( -getRotation() );
rot.multiply( *this );
rot.apply(x1, y1);
x = x1-1;
y = y1-1;
}
I think...
My current favourite word is: I'm starting to run out of fav. words!
-SK Genius
Game Programming articles start - here[ ^]-
|
|
|
|
|
Jim,
SK is right, you should just be able to get these from inspection. Perhaps I don't understand what you want to extract. Do you want the shear equations?
|
|
|
|
|
|
Hi Guys. I posted this in the C# forum but I can't seem to get any help with finding something that can return a 16 digit number. I have a list of numbers that I need to sum up but the problem is that it totals to a 16 digit number or more. Problem is that currently I can't find anything that can do this in C# unless somebody here knows of anything else I can use.
Does anybody have any ideas??
Thanks in advance
Excellence is doing ordinary things extraordinarily well.
|
|
|
|
|
If what you're totaling is overflowing a 16-bit number then you cannot return a 16 bit number without losing precision. You'd have to return at least a 32 bit number with the result of the summation.
For example:
short[] numbers = new short[] { short.MaxValue, short.MaxValue, short.MaxValue, short.MaxValue };
short total = 0;
foreach (short number in numbers)
total += number;
Console.WriteLine(total);
Returns -4 as the result, obviously not what you want, this is because you've overflowed and the number has rolled back to zero. If you change total to an int then you get the correct result.
I doubt it. If it isn't intuitive then we need to fix it. - Chris Maunder
|
|
|
|
|
Well, a 16-digit number is a 16-bit one only in base 2.
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke
|
|
|
|
|
I doubt it. If it isn't intuitive then we need to fix it. - Chris Maunder
|
|
|
|
|
You could just make your own class that deals with the numbers in binary?
|
|
|
|
|
I'm not a C# user, but a few moments with google threw up:
Int64[^]
and
Big Integer Class[^]
Peter
"Until the invention of the computer, the machine gun was the device that enabled humans to make the most mistakes in the smallest amount of time."
|
|
|
|
|
UInt64 tops out at 2^64, which is around 20 digits in base 10.
Theres an article on the MSDN on "integral types in .Net" (from the top of my head) which gives all the details. Decimal is quite a bit bigger, similar to the SQL currency type. Decimal is usually large enough for most purposes.
- Anyone with enough things that you need bigger than decimal to count them can afford to have a bigint library written for them 
|
|
|
|
|
|
If you can't find the equivalent of int__64 (ULONGLONG for me) in C# you can break your sum into a set of essentially 'multi-digit' digits.
That is: an accumulator for 0-999, another for 1,000-999,000, another for 1,000,000-999,000,000, etc. Now, every time your 0-999 accumulator goes greater than 1000, subtract the 1,000's off that accum and add 1 to the 1,000-999,000 accum.
Do you see it?
Tadeusz Westawic
An ounce of Clever is worth a pound of Experience.
|
|
|
|
|
Hi everyone,
I have a small question, for my school project i need to make a program for counting brick in image.
So I have a picture with nothing but bricks on it, and i need to count them,
the idea i have is to use Sobel algorithm, blur, binaryzation and region search, something like that.
Is is the right way to go, or is there any better one?
Thank you all for your help!
|
|
|
|
|
I would say you have it spot on... filter the image and then apply a region counting algorithm.
The filtering is relatively simple... but the region finding could be harder (I've never used/made one).
If you get the original image down to just the outlines of the bricks (via the filters), then you could use a Monte-Carlo based search to look for borders surrounding each point.
[I wish I had a school project as interesting as yours!]
Matthew Butler
|
|
|
|
|
Hmm ya all i need is some guidance where i can get this algorithms implemented in C#.
Thnx for your reply.
|
|
|
|
|
Member 3897207 wrote: all i need is some guidance where i can get this algorithms implemented in C#.
First, you ask for some advice on if your approach to the problem, which is fine. But now you respond to Mathew's advice saying you just want help on how to cheat to do your homework. I was going add a litte to his advice but now I won't and you'll probably not get any further help.
2 75 22 6
|
|
|
|
|
Well what can i say...basically this came out wrong, because i found some algorithms, but they are written in java, so i wondered if there is any...ok...not just for this problem, but if there is any site with algorithms (general) where code is in C#.
Well if i don't get any help any more...so be it.
Thnx anyway at least for confirming that the approach is at least a bit correct.
|
|
|
|
|
C# is very similar to Java (C# partly originated from it). You should be able to translate the code with little hassle. If you get stuck, post your code, and so long as your question is reasonable, we'd be happy to help.
|
|
|
|
|
Yes, to echo what Mark said. The way to approach us on this is to present the problem, describe what you've done to solve it, and ask why you're not getting the results you expect. Then you'll probably get some help. But to post like you're simply looking for a place to cut and paste your homework won't fly. We get plenty of queries like that and they're generally all told to take a hike, some less politely than others.
One potential problem with the method you sketched is that the lines you get after you run the Sobel and threshold is they may contain gaps for one reason or another. You may need a way to look at the partial lines and fill in the blanks. Since the problem is fairly artificial, you have a lot of prior knowledge about what you're looking for.
2 75 22 6
|
|
|
|
|
Tim Craig wrote: to post like you're simply looking for a place to cut and paste your homework won't fly
I ditto that Tim. I was going to give some input, but not now 
"The clue train passed his station without stopping." - John Simmons / outlaw programmer
"Real programmers just throw a bunch of 1s and 0s at the computer to see what sticks" - Pete O'Hanlon
|
|
|
|
|
Well project done...basically I managed to get all algorithems done, except region search, but well in the end I used AForge .NET Framework and got the job done in the fraction of time.
Thank you all for your help and assistance.
|
|
|
|
|
Hi to all,
I want to send a video file on network (UDP) in C++. But when I do that, it doesn't receive properly.
So can anyone help me or give me an algorithm/code to send data on IP continously at a particular bitrate in C++?
Thanks & Regards,
Aniket A. Salunkhe
|
|
|
|
|
Hi
I work with some very large arrays (millions of items) in an algorithm, however, since most fields contains the value 0 I decided to use a Dictionary<int, int=""> to represent only the array's fields where the value is not 0.
Dictionary<int, int=""> means that the first int is the index of the array in which the value would appear, and the second int is the value of this index.
Now, I have two of these "arrays" that I need to compare. For each _existing_ index I need to find the difference between the arrays.
E.g. index i exists in dict1 (first dictionary) but not in dict2, thus the difference should be the value of dict1[i] alone. index i exists in both dictionaries thus the difference should be dict1[i] - dict2[i].
Now, in theory this should be possible to achieve with iterating each unique index in BOTH dictionaries only once (e.g. if an index occurs in both dictionaries it should only be iterated once)
My suggestion was that I start by getting the first item from each and compare the index of these.if they are equal I would increase both iterators.if one is higher than the other I would increase the iterator until they are equal or the other one is lower. This would be repeated until both are iterated through. Any ideas how I can do this with C# and Dictionary<int, int="">?
Any ideas? need to do this iteration as fast as possible. In C I could do some pointer magic to achieve this, but I am not familiar with C# dictionary iterators/enumerators so I have no idea whatsoever how to solve this.
modified on Friday, May 9, 2008 2:50 PM
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
