We Keep Coding

Plot values of a line from a grayscale image

I am trying to take a grayscale bitmap and extract a single line from it and then graph the gray values. I got something to work, but I'm not really happy with it. It just seems slow and tedious. I am sure someone has a better idea
WriteableBitmap someImg; //camera image
int imgWidth = someImg.PixelWidth;
int imgHeight = someImg.PixelHeight;
Int32Rect rectLine = new Int32Rect(0, imgHeight / 2, imgWidth, 1); //horizontal line half way down the image as a rectangle with height 1
//calculate stride and buffer size
int imgStride = (imgWidth * someImg.Format.BitsPerPixel + 7) / 8; // not sure I understand this part
byte[] buffer = new byte[imgStride * rectLine.Height];
//copy pixels to buffer
someImg.CopyPixels(rectLine, buffer, imgStride, 0);
const int xGraphHeight = 256;
WriteableBitmap xgraph = new WriteableBitmap(imgWidth, xGraphHeight, someImg.DpiX, someImg.DpiY, PixelFormats.Gray8, null);
//loop through pixels
for (int i = 0; i < imgWidth; i++)
{
Int32Rect dot = new Int32Rect(i, buffer[i], 1, 1); //1x1 rectangle
byte[] WhiteDotByte = { 255 }; //white
xgraph.WritePixels(dot, WhiteDotByte, imgStride, 0);//write pixel
}
You can see the image and the plot below the green line. I guess I am having some WPF issues that make it look funny but that's a problem for another post.

I assume the goal is to create a plot of the pixel value intensities of the selected line.
The first approach to consider it to use an actual plotting library. I have used oxyplot, it works fine, but is lacking in some aspects. Unless you have specific performance requirements this will likely be the most flexible approach to take.
If you actually want to render to an image you might be better of using unsafe code to access the pixel values directly. For example:
xgraph.Lock();
for (int y = 0; y < imgHeight; y++){
var rowPtr = (byte*)(xgraph.BackBuffer + y * xgraph.BackBufferStride);
for(int x = 0; x < imgWidth; x++){
rowPtr[x] = (byte)(y < buffer[i] ? 0 : 255);
}
}
self.Unlock(); // this should be placed in a finally statement
This should be faster than writing 1x1 rectangles. It should also write columns instead of single pixels, and that should help making the graph more visible. You might also consider allowing arbitrary image height and scale the comparison value.
If you want to plot the pixel values along an arbitrary line, and not just a horizontal one. You can take equidistant samples along the line, and use bilinear interpolation to sample the image.

Faster Image Processing than Lock Bits

I've been working on an edge detection program in C#, and to make it run faster, I recently made it use lock bits. However, lockBits is still not as fast as I would like it to run. Although the problem could be my general algorithm, I'm also wondering if there is anything better than lockBits I can use for image processing.
In case the problem is the algorithm, here's a basic explanation. Go through an array of Colors (made using lockbits, which represent pixels) and for each Color, check the color of the eight pixels around that pixel. If those pixels do not match the current pixel closely enough, consider the current pixel an edge.
Here's the basic code that defines if a pixel is an edge. It takes in a Color[] of nine colors, the first of which is the pixel is to check.
public Boolean isEdgeOptimized(Color[] colors)
{
//colors[0] should be the checking pixel
Boolean returnBool = true;
float percentage = percentageInt; //the percentage used is set
//equal to the global variable percentageInt
if (isMatching(colors[0], colors[1], percentage) &&
isMatching(colors[0], colors[2], percentage) &&
isMatching(colors[0], colors[3], percentage) &&
isMatching(colors[0], colors[4], percentage) &&
isMatching(colors[0], colors[5], percentage) &&
isMatching(colors[0], colors[6], percentage) &&
isMatching(colors[0], colors[7], percentage) &&
isMatching(colors[0], colors[8], percentage))
{
returnBool = false;
}
return returnBool;
}
This code is applied for every pixel, the colors of which are fetched using lockbits.
So basically, the question is, how can I get my program to run faster? Is it my algorithm, or is there something I can use that is faster than lockBits?
By the way, the project is on gitHub, here

Are you really passing in a floating point number as a percentage to isMatching?
I looked at your code for isMatching on GitHub and well, yikes. You ported this from Java, right? C# uses bool not Boolean and while I don't know for sure, I don't like the looks of code that does that much boxing and unboxing. Further, you're doing a ton of floating point multiplication and comparison when you don't need to:
public static bool IsMatching(Color a, Color b, int percent)
{
//this method is used to identify whether two pixels,
//of color a and b match, as in they can be considered
//a solid color based on the acceptance value (percent)
int thresh = (int)(percent * 255);
return Math.Abs(a.R - b.R) < thresh &&
Math.Abs(a.G - b.G) < thresh &&
Math.Abs(a.B - b.B) < thresh;
}
This will cut down the amount of work you're doing per pixel. I still don't like it because I try to avoid method calls in the middle of a per-pixel loop especially an 8x per-pixel loop. I made the method static to cut down on an instance being passed in that isn't used. These changes alone will probably double your performance since we're doing only 1 multiply, no boxing, and are now using the inherent short-circuit of && to cut down the work.
If I were doing this, I'd be more likely to do something like this:
// assert: bitmap.Height > 2 && bitmap.Width > 2
BitmapData data = bitmap.LockBits(new Rectangle(0, 0, bitmap.Width, bitmap.Height),
ImageLockMode.ReadWrite, PixelFormat.Format24bppRgb);
int scaledPercent = percent * 255;
unsafe {
byte* prevLine = (byte*)data.Scan0;
byte* currLine = prevLine + data.Stride;
byte* nextLine = currLine + data.Stride;
for (int y=1; y < bitmap.Height - 1; y++) {
byte* pp = prevLine + 3;
byte* cp = currLine + 3;
byte* np = nextLine + 3;
for (int x = 1; x < bitmap.Width - 1; x++) {
if (IsEdgeOptimized(pp, cp, np, scaledPercent))
{
// do what you need to do
}
pp += 3; cp += 3; np += 3;
}
prevLine = currLine;
currLine = nextLine;
nextLine += data.Stride;
}
}
private unsafe static bool IsEdgeOptimized(byte* pp, byte* cp, byte* np, int scaledPecent)
{
return IsMatching(cp, pp - 3, scaledPercent) &&
IsMatching(cp, pp, scaledPercent) &&
IsMatching(cp, pp + 3, scaledPercent) &&
IsMatching(cp, cp - 3, scaledPercent) &&
IsMatching(cp, cp + 3, scaledPercent) &&
IsMatching(cp, np - 3, scaledPercent) &&
IsMatching(cp, np, scaledPercent) &&
IsMatching(cp, np + 3, scaledPercent);
}
private unsafe static bool IsMatching(byte* p1, byte* p2, int thresh)
{
return Math.Abs(p1++ - p2++) < thresh &&
Math.Abs(p1++ - p2++) < thresh &&
Math.Abs(p1 - p2) < thresh;
}
Which now does all kinds of horrible pointer mangling to cut down on array accesses and so on. If all of this pointer work makes you feel uncomfortable, you can allocate byte arrays for prevLine, currLine and nextLine and do a Marshal.Copy for each row as you go.
The algorithm is this: start one pixel in from the top and left and iterate over every pixel in the image except the outside edge (no edge conditions! Yay!). I keep pointers to the starts of each line, prevLine, currLine, and nextLine. Then when I start the x loop, I make up pp, cp, np which are previous pixel, current pixel and next pixel. current pixel is really the one we care about. pp is the pixel directly above it, np directly below it. I pass those into IsEdgeOptimized which looks around cp, calling IsMatching for each.
Now this all assume 24 bits per pixel. If you're looking at 32 bits per pixel, all those magic 3's in there need to be 4's, but other than that the code doesn't change. You could parameterize the number of bytes per pixel if you want so it could handle either.
FYI, the channels in the pixels are typically b, g, r, (a).
Colors are stored as bytes in memory. Your actual Bitmap, if it is a 24 bit image is stored as a block of bytes. Scanlines are data.Stride bytes wide, which is at least as large as 3 * the number of pixels in a row (it may be larger because scan lines are often padded).
When I declare a variable of type byte * in C#, I'm doing a few things. First, I'm saying that this variable contains the address of a location of a byte in memory. Second, I'm saying that I'm about to violate all the safety measures in .NET because I could now read and write any byte in memory, which can be dangerous.
So when I have something like:
Math.Abs(*p1++ - *p2++) < thresh
What it says is (and this will be long):
Take the byte that p1 points to and hold onto it
Add 1 to p1 (this is the ++ - it makes the pointer point to the next byte)
Take the byte that p2 points to and hold onto it
Add 1 to p2
Subtract step 3 from step 1
Pass that to Math.Abs.
The reasoning behind this is that, historically, reading the contents of a byte and moving forward is a very common operation and one that many CPUs build into a single operation of a couple instructions that pipeline into a single cycle or so.
When we enter IsMatching, p1 points to pixel 1, p2 points to pixel 2 and in memory they are laid out like this:
p1 : B
p1 + 1: G
p1 + 2: R
p2 : B
p2 + 1: G
p2 + 2: R
So IsMatching just does the the absolute difference while stepping through memory.
Your follow-on question tells me that you don't really understand pointers. That's OK - you can probably learn them. Honestly, the concepts really aren't that hard, but the problem with them is that without a lot of experience, you are quite likely to shoot yourself in the foot, and perhaps you should consider just using a profiling tool on your code and cooling down the worst hot spots and call it good.
For example, you'll note that I look from the first row to the penultimate row and the first column to the penultimate column. This is intentional to avoid having to handle the case of "I can't read above the 0th line", which eliminates a big class of potential bugs which would involve reading outside a legal memory block, which may be benign under many runtime conditions.

Instead of copying each image to a byte[], then copying to a Color[], creating another temp Color[9] for each pixel, and then using SetPixel to set the color, compile using the /unsafe flag, mark the method as unsafe, replace copying to a byte[] with Marshal.Copy to:
using (byte* bytePtr = ptr)
{
//code goes here
}
Make sure you replace the SetPixel call with setting the proper bytes. This isn't an issue with LockBits, you need LockBits, the issue is that you're being inefficient with everything else related to processing the image.

If you want to use parallel task execution, you can use the Parallel class in System.Threading.Tasks namespace. Following link has some samples and explanations.
http://csharpexamples.com/fast-image-processing-c/
http://msdn.microsoft.com/en-us/library/dd460713%28v=vs.110%29.aspx

You can split the image into 10 bitmaps and process each one, then finally combine them (just an idea).

Is there an efficient algorithm for segmentation of handwritten text?

I want to automatically divide an image of ancient handwritten text by lines (and by words in future).
The first obvious part is preprocessing the image...
I'm just using a simple digitization (based on brightness of pixel). After that I store data into two-dimensional array.
The next obvious part is analyzing the binary array.
My first algorithm was pretty simple - if there are more black pixels in a row of the array than the root-mean-square of Maximum and Minimum value, then this row is part of line.
After forming the list of lines I cut off lines with height that is less than average.
Finally it turned out into some kind of linear regression, trying to minimize the difference between the blank rows and text rows. (I assumed that fact)
My second attempt - I tried to use GA with several fitness functions.
The chromosome contained 3 values - xo, x1, x2. xo [-1;0] x1 [0;0.5] x2 [0;0.5]
Function, that determines identity the row to line is (xo + α1 x1 + α2 x2) > 0, where α1 is scaled sum of black pixels in row, α2 is median value of ranges between the extreme black pixels in row. (a1,a2 [0,1])
Another functions, that I tried is (x1 < α1 OR x2 > α2) and (1/xo + [a1 x1] / [a2 x2] ) > 0
The last function is the most efficient.
The fitness function is
(1 / (HeigthRange + SpacesRange)
Where range is difference between maximum and minimum. It represents the homogeneity of text. The global optimum of this function - the most smooth way to divide the image into lines.
I am using C# with my self-coded GA (classical, with 2-point crossover, gray-code chromosomes, maximum population is 40, mutation rate is 0.05)
Now I ran out of ideas how to divide this image into lines with ~100% accuracy.
What is the efficient algorithm to do this?
UPDATE:
Original BMP (1.3 MB)
UPDATE2:
Improved results on this text to 100%
How I did it:
fixed minor bug in range count
changed fitness function to 1/(distancesRange+1)*(heightsRange+1))
minimized classifying function to (1/xo + x2/range) > 0 (points in row now don't affect classification)
(i.e. optimized input data and made fitness function optimizations more explicit)
Problem:
GA surprisingly failed to recognize this line. I looked at debug data of 'find rages' function and found, that there is too much noise in 'unrecognized' place.
The function code is below:
public double[] Ranges()
{
var ranges = new double[_original.Height];
for (int y = 0; y < _original.Height; y++ )
{
ranges[y] = 0;
var dx = new List<int>();
int last = 0;
int x = 0;
while (last == 0 && x<_original.Width)
{
if (_bit[x, y])
last = x;
x++;
}
if (last == 0)
{
ranges[y] = 0;
continue;
}
for (x = last; x<_original.Width; x++)
{
if (!_bit[x, y]) continue;
if (last != x - 1)
{
dx.Add((x-last)+1);
}
last = x;
}
if (dx.Count > 2)
{
dx.Sort();
ranges[y] = dx[dx.Count / 2];
//ranges[y] = dx.Average();
}
else
ranges[y] = 0;
}
var maximum = ranges.Max();
for (int i = 0; i < ranges.Length; i++)
{
if (Math.Abs(ranges[i] - 0) < 0.9)
ranges[i] = maximum;
}
return ranges;
}
I'm using some hacks in this code. The main reason - I want to minimize the range between nearest black pixels, but if there are no pixels, the value becomes '0', and it becomes impossible to solve this problem with finding optimas. The second reason - this code is changing too frequently.
I'll try to fully change this code, but I have no idea how to do it.
Q:
If there is more efficient fitness function?
How to find more versatile determination function?

Although I'm not sure how to translate the following algorithm into GA (and I'm not sure why you need to use GA for this problem), and I could be off base in proposing it, here goes.
The simple technique I would propose is to count the number of black pixels per row. (Actually it's the dark pixel density per row.) This requires very few operations, and with a few additional calculations it's not difficult to find peaks in the pixel-sum histogram.
A raw histogram will look something like this, where the profile along the left side shows the number of dark pixels in a row. For visibility, the actual count is normalized to stretch out to x = 200.
After some additional, simple processing is added (described below), we can generate a histogram like this that can be clipped at some threshold value. What remains are peaks indicating the center of lines of text.
From there it's a simple matter to find the lines: just clip (threshold) the histogram at some value such as 1/2 or 2/3 the maximum, and optionally check that the width of the peak at your clipping threshold is some minimum value w.
One implementation of the full (yet still simple!) algorithm to find the nicer histogram is as follows:
Binarize the image using a "moving average" threshold or similar local thresholding technique in case a standard Otsu threshold operating on pixels near edges isn't satisfactory. Or, if you have a nice black-on-white image, just use 128 as your binarization threshold.
Create an array to store your histogram. This array's length will be the height of the image.
For each pixel (x,y) in the binarized image, find the number of dark pixels above and below (x,y) at some radius R. That is, count the number of dark pixels from (x, y - R) to x (y + R), inclusive.
If the number of dark pixels within a vertical radius R is equal or greater to R--that is, at least half the pixels are dark--then pixel (x,y) has sufficient vertical dark neighbors. Increment your bin count for row y.
As you march along each row, track the leftmost and rightmost x-values for pixels with sufficient neighbors. As long as the width (right - left + 1) exceeds some minimum value, divide the total count of dark pixels by this width. This normalizes the count to ensure the short lines like the very last line of text are included.
(Optional) Smooth the resulting histogram. I just used the mean over 3 rows.
The "vertical count" (step 3) eliminates horizontal strokes that happen to be located above or below the center line of text. A more sophisticated algorithm would just check directly above and below (x,y), but also to the upper left, upper right, lower left, and lower right.
With my rather crude implementation in C# I was able to process the image in less than 75 milliseconds. In C++, and with some basic optimization, I've little doubt the time could be cut down considerably.
This histogram method assumes the text is horizontal. Since the algorithm is reasonably fast, you may have enough time to calculate pixel count histograms at increments of every 5 degrees from the horizontal. The scan orientation with the greatest peak/valley differences would indicate the rotation.
I'm not familiar with GA terminology, but if what I've suggested is of some value I'm sure you can translate it into GA terms. In any case, I was interested in this problem anyway, so I might as well share.
EDIT: maybe for use GA, it's better to think in terms of "distance since previous dark pixel in X" (or along angle theta) and "distance since previous dark pixel in Y" (or along angle [theta - pi/2]). You might also check distance from white pixel to dark pixel in all radial directions (to find loops).
byte[,] arr = get2DArrayFromBitamp(); //source array from originalBitmap
int w = arr.GetLength(0); //width of 2D array
int h = arr.GetLength(1); //height of 2D array
//we can use a second 2D array of dark pixels that belong to vertical strokes
byte[,] bytes = new byte[w, h]; //dark pixels in vertical strokes
//initial morph
int r = 4; //radius to check for dark pixels
int count = 0; //number of dark pixels within radius
//fill the bytes[,] array only with pixels belonging to vertical strokes
for (int x = 0; x < w; x++)
{
//for the first r rows, just set pixels to white
for (int y = 0; y < r; y++)
{
bytes[x, y] = 255;
}
//assume pixels of value < 128 are dark pixels in text
for (int y = r; y < h - r - 1; y++)
{
count = 0;
//count the dark pixels above and below (x,y)
//total range of check is 2r, from -r to +r
for (int j = -r; j <= r; j++)
{
if (arr[x, y + j] < 128) count++;
}
//if half the pixels are dark, [x,y] is part of vertical stroke
bytes[x, y] = count >= r ? (byte)0 : (byte)255;
}
//for the last r rows, just set pixels to white
for (int y = h - r - 1; y < h; y++)
{
bytes[x, y] = 255;
}
}
//count the number of valid dark pixels in each row
float max = 0;
float[] bins = new float[h]; //normalized "dark pixel strength" for all h rows
int left, right, width; //leftmost and rightmost dark pixels in row
bool dark = false; //tracking variable
for (int y = 0; y < h; y++)
{
//initialize values at beginning of loop iteration
left = 0;
right = 0;
width = 100;
for (int x = 0; x < w; x++)
{
//use value of 128 as threshold between light and dark
dark = bytes[x, y] < 128;
//increment bin if pixel is dark
bins[y] += dark ? 1 : 0;
//update leftmost and rightmost dark pixels
if (dark)
{
if (left == 0) left = x;
if (x > right) right = x;
}
}
width = right - left + 1;
//for bins with few pixels, treat them as empty
if (bins[y] < 10) bins[y] = 0;
//normalize value according to width
//divide bin count by width (leftmost to rightmost)
bins[y] /= width;
//calculate the maximum bin value so that bins can be scaled when drawn
if (bins[y] > max) max = bins[y];
}
//calculated the smoothed value of each bin i by averaging bin i-1, i, and i+1
float[] smooth = new float[bins.Length];
smooth[0] = bins[0];
smooth[smooth.Length - 1] = bins[bins.Length - 1];
for (int i = 1; i < bins.Length - 1; i++)
{
smooth[i] = (bins[i - 1] + bins[i] + bins[i + 1])/3;
}
//create a new bitmap based on the original bitmap, then draw bins on top
Bitmap bmp = new Bitmap(originalBitmap);
using (Graphics gr = Graphics.FromImage(bmp))
{
for (int y = 0; y < bins.Length; y++)
{
//scale each bin so that it is drawn 200 pixels wide from the left edge
float value = 200 * (float)smooth[y] / max;
gr.DrawLine(Pens.Red, new PointF(0, y), new PointF(value, y));
}
}
pictureBox1.Image = bmp;

After fiddling around this for a while I found that I simply need to count the number of crossings for each line, that is, a switch from white to black would count as one, and a switch from black to white would increment by one again. By highlighting each line with a count > 66 I got close to 100% accuracy, except for the bottom most line.
Of course, would not be robust to slightly rotated scanned documents. And there is this disadvantage of needing to determine the correct threshold.

IMHO with the image shown that would be so hard to do 100% perfectly.
My answer is to give you alternate idea's.
Idea 1:
Make your own version of ReCaptcha (to put on your very own pron site) - and make it a fun game.. "Like cut out a word (edges should all be white space - with some tolerance for overlapping chars on above and below lines)."
Idea 2:
This was a game we played as kids, the wire of a coat hanger was all bent in waves and connected to a buzzer and you had to navigate a wand with a ring in the end with the wire through it, across one side to the other without making the buzzer go off. Perhaps you could adapt this idea and make a mobile game where people trace out the lines without touching black text (with tolerance for overlapping chars)... when they can do a line they get points and get to new levels where you give them harder images..
Idea 3:
Research how google/recaptcha got around it
Idea 4:
Get the SDK for photoshop and master the functionality of it Extract Edges tool
Idea 5:
Stretch the image heaps on the Y Axis which should help, apply the algorithm, then reduce the location measurements and apply them on the normal sized image.

Hough Circle in c# , dynamic memory

I have problem in dynamic memory allocation getting data from matrix
Image image_gray = new Image("im1.jpg");
Matrix circles = new Matrix(100, 1, 3);
Question 1: How can I locate dynamic memory because I don't know the number of circles?
Emgu.CV.CvInvoke.cvHoughCircles(image_gray, circles, HOUGH_TYPE.CV_HOUGH_GRADIENT,
2, 100, 200, 100, 10, 500);
Question 2: Now circle is matrix with [100 ,3], How can I get
point center= Round (circle[i][1], circle[i][1])
How can I get
int radius= circle[i][2];
what should my for loop look like to get data from matrix and casting should be point and int.
I tried already (NOT WORKING / ERROR)
for (int i=0; i < circles.Rows; i++)
{ Matrix entry = circles.GetRow(i);
float x = entry[0];
float y = entry[1];
float r = entry[2];} // NOT WORKING
because instead of copying (required) only (i) row it copies whole matrix( circles) and float x=......gives errors
No overload for method 'this' takes '1' arguments
please help me in this regard
regards
sorry to answer my question. please give some hint to dynamic memory allocation and matrix data

The documentation suggests that it will grow the matrix to fit, but it also contradicts that, so honestly, I would try a 1 row 1 col matrix on an image with more circles and see what it gives you. I see that you're using Matrix type - I don't know how you convert to the IntPtr needed by cvHoughCircles, but it should be straight forward to iterate over each row and pull out the results, which are 3 floats: (x, y, r) -
for (int i=0; i < circles.Cols; i++) {
Matrix<float> entry = circles.GetRow(i);
float x = entry[0]; // guessing here - the doc'n is truly awful
float y = entry[1];
float r = entry[2];
}

Efficient ways to determine tilt of an image

I'm trying to write a program to programmatically determine the tilt or angle of rotation in an arbitrary image.
Images have the following properties:
Consist of dark text on a light background
Occasionally contain horizontal or vertical lines which only intersect at 90 degree angles.
Skewed between -45 and 45 degrees.
See this image as a reference (its been skewed 2.8 degrees).
So far, I've come up with this strategy: Draw a route from left to right, always selecting the nearest white pixel. Presumably, the route from left to right will prefer to follow the path between lines of text along the tilt of the image.
Here's my code:
private bool IsWhite(Color c) { return c.GetBrightness() >= 0.5 || c == Color.Transparent; }
private bool IsBlack(Color c) { return !IsWhite(c); }
private double ToDegrees(decimal slope) { return (180.0 / Math.PI) * Math.Atan(Convert.ToDouble(slope)); }
private void GetSkew(Bitmap image, out double minSkew, out double maxSkew)
{
decimal minSlope = 0.0M;
decimal maxSlope = 0.0M;
for (int start_y = 0; start_y < image.Height; start_y++)
{
int end_y = start_y;
for (int x = 1; x < image.Width; x++)
{
int above_y = Math.Max(end_y - 1, 0);
int below_y = Math.Min(end_y + 1, image.Height - 1);
Color center = image.GetPixel(x, end_y);
Color above = image.GetPixel(x, above_y);
Color below = image.GetPixel(x, below_y);
if (IsWhite(center)) { /* no change to end_y */ }
else if (IsWhite(above) && IsBlack(below)) { end_y = above_y; }
else if (IsBlack(above) && IsWhite(below)) { end_y = below_y; }
}
decimal slope = (Convert.ToDecimal(start_y) - Convert.ToDecimal(end_y)) / Convert.ToDecimal(image.Width);
minSlope = Math.Min(minSlope, slope);
maxSlope = Math.Max(maxSlope, slope);
}
minSkew = ToDegrees(minSlope);
maxSkew = ToDegrees(maxSlope);
}
This works well on some images, not so well on others, and its slow.
Is there a more efficient, more reliable way to determine the tilt of an image?

I've made some modifications to my code, and it certainly runs a lot faster, but its not very accurate.
I've made the following improvements:
Using Vinko's suggestion, I avoid GetPixel in favor of working with bytes directly, now the code runs at the speed I needed.
My original code simply used "IsBlack" and "IsWhite", but this isn't granular enough. The original code traces the following paths through the image:
http://img43.imageshack.us/img43/1545/tilted3degtextoriginalw.gif
Note that a number of paths pass through the text. By comparing my center, above, and below paths to the actual brightness value and selecting the brightest pixel. Basically I'm treating the bitmap as a heightmap, and the path from left to right follows the contours of the image, resulting a better path:
http://img10.imageshack.us/img10/5807/tilted3degtextbrightnes.gif
As suggested by Toaomalkster, a Gaussian blur smooths out the height map, I get even better results:
http://img197.imageshack.us/img197/742/tilted3degtextblurredwi.gif
Since this is just prototype code, I blurred the image using GIMP, I did not write my own blur function.
The selected path is pretty good for a greedy algorithm.
As Toaomalkster suggested, choosing the min/max slope is naive. A simple linear regression provides a better approximation of the slope of a path. Additionally, I should cut a path short once I run off the edge of the image, otherwise the path will hug the top of the image and give an incorrect slope.
Code
private double ToDegrees(double slope) { return (180.0 / Math.PI) * Math.Atan(slope); }
private double GetSkew(Bitmap image)
{
BrightnessWrapper wrapper = new BrightnessWrapper(image);
LinkedList<double> slopes = new LinkedList<double>();
for (int y = 0; y < wrapper.Height; y++)
{
int endY = y;
long sumOfX = 0;
long sumOfY = y;
long sumOfXY = 0;
long sumOfXX = 0;
int itemsInSet = 1;
for (int x = 1; x < wrapper.Width; x++)
{
int aboveY = endY - 1;
int belowY = endY + 1;
if (aboveY < 0 || belowY >= wrapper.Height)
{
break;
}
int center = wrapper.GetBrightness(x, endY);
int above = wrapper.GetBrightness(x, aboveY);
int below = wrapper.GetBrightness(x, belowY);
if (center >= above && center >= below) { /* no change to endY */ }
else if (above >= center && above >= below) { endY = aboveY; }
else if (below >= center && below >= above) { endY = belowY; }
itemsInSet++;
sumOfX += x;
sumOfY += endY;
sumOfXX += (x * x);
sumOfXY += (x * endY);
}
// least squares slope = (NΣ(XY) - (ΣX)(ΣY)) / (NΣ(X^2) - (ΣX)^2), where N = elements in set
if (itemsInSet > image.Width / 2) // path covers at least half of the image
{
decimal sumOfX_d = Convert.ToDecimal(sumOfX);
decimal sumOfY_d = Convert.ToDecimal(sumOfY);
decimal sumOfXY_d = Convert.ToDecimal(sumOfXY);
decimal sumOfXX_d = Convert.ToDecimal(sumOfXX);
decimal itemsInSet_d = Convert.ToDecimal(itemsInSet);
decimal slope =
((itemsInSet_d * sumOfXY) - (sumOfX_d * sumOfY_d))
/
((itemsInSet_d * sumOfXX_d) - (sumOfX_d * sumOfX_d));
slopes.AddLast(Convert.ToDouble(slope));
}
}
double mean = slopes.Average();
double sumOfSquares = slopes.Sum(d => Math.Pow(d - mean, 2));
double stddev = Math.Sqrt(sumOfSquares / (slopes.Count - 1));
// select items within 1 standard deviation of the mean
var testSample = slopes.Where(x => Math.Abs(x - mean) <= stddev);
return ToDegrees(testSample.Average());
}
class BrightnessWrapper
{
byte[] rgbValues;
int stride;
public int Height { get; private set; }
public int Width { get; private set; }
public BrightnessWrapper(Bitmap bmp)
{
Rectangle rect = new Rectangle(0, 0, bmp.Width, bmp.Height);
System.Drawing.Imaging.BitmapData bmpData =
bmp.LockBits(rect,
System.Drawing.Imaging.ImageLockMode.ReadOnly,
bmp.PixelFormat);
IntPtr ptr = bmpData.Scan0;
int bytes = bmpData.Stride * bmp.Height;
this.rgbValues = new byte[bytes];
System.Runtime.InteropServices.Marshal.Copy(ptr,
rgbValues, 0, bytes);
this.Height = bmp.Height;
this.Width = bmp.Width;
this.stride = bmpData.Stride;
}
public int GetBrightness(int x, int y)
{
int position = (y * this.stride) + (x * 3);
int b = rgbValues[position];
int g = rgbValues[position + 1];
int r = rgbValues[position + 2];
return (r + r + b + g + g + g) / 6;
}
}
The code is good, but not great. Large amounts of whitespace cause the program to draw relatively flat line, resulting in a slope near 0, causing the code to underestimate the actual tilt of the image.
There is no appreciable difference in the accuracy of the tilt by selecting random sample points vs sampling all points, because the ratio of "flat" paths selected by random sampling is the same as the ratio of "flat" paths in the entire image.

GetPixel is slow. You can get an order of magnitude speed up using the approach listed here.

If text is left (right) aligned you can determine the slope by measuring the distance between the left (right) edge of the image and the first dark pixel in two random places and calculate the slope from that. Additional measurements would lower the error while taking additional time.

First I must say I like the idea. But I've never had to do this before and I'm not sure what all to suggest to improve reliability. The first thing I can think of this is this idea of throwing out statistical anomalies. If the slope suddenly changes sharply then you know you've found a white section of the image that dips into the edge skewing (no pun intended) your results. So you'd want to throw that stuff out somehow.
But from a performance standpoint there are a number of optimizations you could make which may add up.
Namely, I'd change this snippet from your inner loop from this:
Color center = image.GetPixel(x, end_y);
Color above = image.GetPixel(x, above_y);
Color below = image.GetPixel(x, below_y);
if (IsWhite(center)) { /* no change to end_y */ }
else if (IsWhite(above) && IsBlack(below)) { end_y = above_y; }
else if (IsBlack(above) && IsWhite(below)) { end_y = below_y; }
To this:
Color center = image.GetPixel(x, end_y);
if (IsWhite(center)) { /* no change to end_y */ }
else
{
Color above = image.GetPixel(x, above_y);
Color below = image.GetPixel(x, below_y);
if (IsWhite(above) && IsBlack(below)) { end_y = above_y; }
else if (IsBlack(above) && IsWhite(below)) { end_y = below_y; }
}
It's the same effect but should drastically reduce the number of calls to GetPixel.
Also consider putting the values that don't change into variables before the madness begins. Things like image.Height and image.Width have a slight overhead every time you call them. So store those values in your own variables before the loops begin. The thing I always tell myself when dealing with nested loops is to optimize everything inside the most inner loop at the expense of everything else.
Also... as Vinko Vrsalovic suggested, you may look at his GetPixel alternative for yet another boost in speed.

At first glance, your code looks overly naive.
Which explains why it doesn't always work.
I like the approach Steve Wortham suggested,
but it might run into problems if you have background images.
Another approach that often helps with images is to blur them first.
If you blur your example image enough, each line of text will end up
as a blurry smooth line. You then apply some sort of algorithm to
basically do a regression analisys. There's lots of ways to do
that, and lots of examples on the net.
Edge detection might be useful, or it might cause more problems that its worth.
By the way, a gaussian blur can be implemented very efficiently if you search hard enough for the code. Otherwise, I'm sure there's lots of libraries available.
Haven't done much of that lately so don't have any links on hand.
But a search for Image Processing library will get you good results.
I'm assuming you're enjoying the fun of solving this, so not much in actual implementation detalis here.

Measuring the angle of every line seems like overkill, especially given the performance of GetPixel.
I wonder if you would have better performance luck by looking for a white triangle in the upper-left or upper-right corner (depending on the slant direction) and measuring the angle of the hypotenuse. All text should follow the same angle on the page, and the upper-left corner of a page won't get tricked by the descenders or whitespace of content above it.
Another tip to consider: rather than blurring, work within a greatly-reduced resolution. That will give you both the smoother data you need, and fewer GetPixel calls.
For example, I made a blank page detection routine once in .NET for faxed TIFF files that simply resampled the entire page to a single pixel and tested the value for a threshold value of white.

What are your constraints in terms of time?
The Hough transform is a very effective mechanism for determining the skew angle of an image. It can be costly in time, but if you're going to use Gaussian blur, you're already burning a pile of CPU time. There are also other ways to accelerate the Hough transform that involve creative image sampling.

Your latest output is confusing me a little.
When you superimposed the blue lines on the source image, did you offset it a bit? It looks like the blue lines are about 5 pixels above the centre of the text.
Not sure about that offset, but you definitely have a problem with the derived line "drifting" away at the wrong angle. It seems to have too strong a bias towards producing a horizontal line.
I wonder if increasing your mask window from 3 pixels (centre, one above, one below) to 5 might improve this (two above, two below). You'll also get this effect if you follow richardtallent's suggestion and resample the image smaller.

Very cool path finding application.
I wonder if this other approach would help or hurt with your particular data set.
Assume a black and white image:
Project all black pixels to the right (EAST). This should give a result of a one dimensional array with a size of IMAGE_HEIGHT. Call the array CANVAS.
As you project all the pixels EAST, keep track numerically of how many pixels project into each bin of CANVAS.
Rotate the image an arbitrary number of degrees and re-project.
Pick the result that gives the highest peaks and lowest valleys for values in CANVAS.
I imagine this will not work well if in fact you have to account for a real -45 -> +45 degrees of tilt. If the actual number is smaller(?+/- 10 degrees), this might be a pretty good strategy. Once you have an intial result, you could consider re-running with a smaller increment of degrees to fine tune the answer. I might therefore try to write this with a function that accepted a float degree_tick as a parm so I could run both a coarse and fine pass (or a spectrum of coarseness or fineness) with the same code.
This might be computationally expensive. To optimize, you might consider selecting just a portion of the image to project-test-rotate-repeat on.