Develop Reference

C#: Improving edge point extraction algorithm

I am working on an edge points extraction algorithm.
I have a List of points representing a blob (a group of connected pixels) and I want to extract edge points.
I have an example algorithm below, but I am wondering if there is a faster way.
I am using class for BlobPoint, because there is more to it than shown in the example below.
BlobPoint index value represent index value in the original image (image is 1D array of pixels, hence required width information).
This algorithm is build on idea that if a pixel on right or bottom or left or top does not exist in the list, then this point is an edge point.
The list will usually contain between 20 000 to 1 000 000 elements.
yr and yl are added for clarity.
public static List<BlobPoint> GetEdgePoints(this List<BlobPoint> points, int width)
{
int length = points.Count;
List<BlobPoint> temp = new List<BlobPoint>();
for (int i = 0; i < length; i++)
{
BlobPoint point = points[i];
int x = point.X;
int y = point.Y;
int xr = x + 1;
int yr = y;
int xb = x - width;
int yb = y - 1;
int xl = x - 1;
int yl = y;
int xt = x + width;
int yt = y + 1;
if (!points.Any(p => p.X == xb && p.Y == yb) || !points.Any(p => p.X == xl && p.Y == yl) || !points.Any(p => p.X == xr && p.Y == yr) || !points.Any(p => p.X == xt && p.Y == yt))
{
temp.Add(point);
}
}
return temp;
}
public class BlobPoint
{
public int X = 0;
public int Y = 0;
public int Index = 0;
public BlobPoint(int x, int y, int index)
{
X = x;
Y = y;
Index = index;
}
}
My solution:
Ok, I have done some testing and the method above is way too slow for my needs.
I will leave question as this might be of use to someone who looks for faster iteration method.
This is a solution I came up with, but it is still a bit slow:
public static Dictionary<int, List<BlobPoint>> GetEdgePoints(BlobPoint[] points, int[] labels, int image_width, int image_height)
{
int length = labels.Length;
if (image_height * image_width != length) throw new ArgumentException("image_width x image_height does not match labels.Length");
if (length == 0) throw new ArgumentException("label array cannot be empty!");
if (points.Length != length) throw new ArgumentException("points array length cannot be different from labels array length!");
var dict = new Dictionary<int, List<BlobPoint>>();
for (int i = 0; i < length; i++)
{
int label = labels[i];
if (label <= 0) continue;
BlobPoint point = points[i];
int x = point.X;
int y = point.Y;
int width_offset = image_width - 1;
int height_offset = image_height - 1;
if (x > 0 && x < width_offset && y > 0 && y < height_offset)
{
if (labels[i + 1] == label && labels[i - 1] == label && labels[i + image_width] == label && labels[i - image_width] == label)
{
continue;
}
}
if (dict.ContainsKey(label))
dict[label].Add(point);
else
dict.Add(label, new List<BlobPoint>() { point });
}
return dict;
}

Personally, I'd use a two-dimensional Boolean array representing the image for this, in which the indices for the points you have are set to true. In exchange for some memory, and a single loop over the points in advance, this allows lightning-fast checks of whether a point is inside the list, removing all internal iterations for lookups on the points list. There is not a single .Contains or .Any or .All performed here. Just two loops over the main list, and eight very simple checks inside the second loop.
public static List<BlobPoint> GetEdgePoints(this List<BlobPoint> points, Int32 imageWidth, Int32 imageHeight)
{
Boolean[,] pointInList = new Boolean[imageHeight, imageWidth];
foreach (BlobPoint p in points)
pointInList[p.Y, p.X] = true;
List<BlobPoint> edgePoints = new List<BlobPoint>();
Int32 lastX = imageWidth - 1;
Int32 lastY = imageHeight - 1;
foreach (BlobPoint p in points)
{
Int32 x = p.X;
Int32 y = p.Y;
// Image edge is obviously a blob edge too.
// Coordinates checks are completely safe after the edge checks.
if (x == 0 || y == 0 || x == lastX || y == lastY
|| !pointInList[y - 1, x]
|| !pointInList[y, x - 1]
|| !pointInList[y, x + 1]
|| !pointInList[y + 1, x])
edgePoints.Add(p);
}
return edgePoints;
}
This technically works without the image width and height, but then you need to do another loop in advance to get the maximum x and y present in your points, so you can make a Boolean[,] array that can contain all coordinates. If you have that image data, though, it's obviously a lot more efficient to just use it.
Not sure why you bother with the BlobPoint class over the standard Point struct, though. The index in the original array is just p.Y * stride + p.X anyway.

choosing a next point in boundary

I'm trying to apply bound hall algorithm on set of points which are deployed randomly.
I based on the angle between the previous point and the neighborhood points of the current point to choose the next point with minimal angel.
What i'm asking for is in some cases the boundary stop generating any point and I've actually debugged the code many times and couldn't figure out what the problem.
The pic of case and the code are given below.
ArrayList angles = new ArrayList();
for (int s = 0; s < node.SensorNeighbors.Count; s++)
{
Node n = node.SensorNeighbors[s];
if (backboneNodes.Count != 1) // backboneNode is array hase nodes in boundary
{
Node previous = backboneNodes[backboneNodes.Count - 2];
Double Angle = Math.Atan2(previous.y - node.y, previous.x - node.x) - Math.Atan2(n.y - node.y, n.x - node.x);
angles.Add(Angle);
}
if (backboneNodes.Count == 1)
{
Double Angle = Math.Atan2(520 - node.y, node.x - node.x) - Math.Atan2(n.y - node.y, n.x - node.x);
angles.Add(Angle);
}
}
int index = 0;
double minAngle = int.MaxValue;
for (int s = 0; s < angles.Count; s++)
{
if ((double)angles[s] < minAngle)
{
minAngle = (double)angles[s];
index = angles.IndexOf(minAngle);
}
}
Node Nnode = node.SensorNeighbors[index];
int iRadius = (int)Math.Sqrt(Math.Pow(node.x - Nnode.x, 2) + Math.Pow(node.y - Nnode.y, 2));
node.aConnections.Add(node, Nnode, (int)(fTransmissionCost * iRadius / iTransmissionRadius), iReceiveCost, iTransmitterDelay);
node = Nnode;

Randomly generate blocks on a flat map

I'm trying to randomly generate blocks on a flat map and make it so that they don't overlap each other.
I have made a matrix (c# array) of the size of the map (500x500), the blocks have a scale between 1 and 5.
The code works but if a generated block overlaps another one, it is destroyed and not regenerated somewhere else.
Only around 80 of the 1000 blocks I try to generate don't overlap another block.
Here is a picture of the map with around 80 blocks generated, the green squares are blocks
void generateElement(int ratio, int minScale, int maxScale, GameObject g) {
bool elementFound = false;
for (int i = 0; i < ratio * generationDefault; i++) {
GameObject el;
// Randomly generate block size and position
int size = Random.Range(minScale, maxScale + 1);
int x = Random.Range(0, mapSizex + 1 - size);
int y = Random.Range(0, mapSizey + 1 - size);
// Check if there is already an element
for (int j = x; j < x + size; j++)
for (int k = y; k < y + size; k++)
if (map[j][k] != null)
elementFound = true;
if (elementFound)
continue;
else {
el = (GameObject)Instantiate(g, new Vector3(x + (float)size / 2, (float)size / 2, y + (float)size / 2), Quaternion.Euler(0, 0, 0));
el.transform.localScale *= size;
}
// Create element on map array
for (int j = x; j < x + size; j++)
for (int k = y; k < y + size; k++)
if (map[j][k] == null) {
map[j][k] = el.GetComponent<ObjectInterface>();
}
}
}
I thought of 3 possible fixes
I should set the size of the block depending of the place it has.
I should use another randomization algorithm.
I'm not doing this right.
What do you think is the best idea ?
UPDATE
I got the code working much better. I now try to instantiate the blocks multiple times if needed (maximum 5 for the moment) and I fixed the bugs. If there are already many elements on the map, they will not always be instantiated and that's what I wanted, I just have to find the right amount of times it will try to instantiate the block.
I tried instantiating 1280 elements on a 500x500 map. It takes only about 1.5 second and it instantiated 1278/1280 blocks (99.843%).
void generateElement(int ratio, int minScale, int maxScale, GameObject g) {
bool elementFound = false;
int cnt = 0;
// Generate every block
for (int i = 0; i < ratio * generationDefault; i++) {
GameObject el = null;
// Randomly generate block size and position
int size, x, y, tryCnt = 0;
// Try maximum 5 times to generate the block
do {
elementFound = false;
// Randomly set block size and position
size = Random.Range(minScale, maxScale + 1);
x = Random.Range(0, mapSizex + 1 - size);
y = Random.Range(0, mapSizey + 1 - size);
// Check if there is already an element
for (int j = x; j < x + size; j++)
for (int k = y; k < y + size; k++)
if (map[j][k] != null)
elementFound = true;
tryCnt++;
} while (elementFound && tryCnt < 5);
if (tryCnt >= 5 && elementFound) continue;
// Instantiate the block
el = (GameObject)Instantiate(g, new Vector3(x + (float)size / 2, (float)size / 2, y + (float)size / 2), Quaternion.Euler(0, 0, 0));
el.transform.localScale *= size;
// Create element on map array
for (int j = x; j < x + size; j++)
for (int k = y; k < y + size; k++)
if (map[j][k] == null) {
map[j][k] = el.GetComponent<ObjectInterface>();
}
cnt++;
}
print("Instantiated " + cnt + "/" + ratio * generationDefault);
}

This is incredibly difficult to do well.
Here's a quick solution you'll maybe like ... depending on your scene.
actualWidth = 500 //or whatever. assume here is square
// your blocks are up to 5 size
chunkWidth = actualWidth / 5
// it goes without saying, everything here is an int
kChunks = chunkWidth*chunkWidth
List<int> shuf = Enumerable.Range(1,kChunks).OrderBy(r=>Random.value).ToList();
howManyWanted = 1000
shuf = shuf.Take(howManyWanted)
foreach( i in shuf )
x = i % actualWidth
y = i / actualWidth
make block at x y
put block in list allBlocks
HOWEVER ............
...... you'll see that this looks kind of "regular", so do this:
Just randomly perturb all the blocks. Remember, video game programming is about clever tricks!
Ideally, you have to start from the middle and work your way out; in any event you can't just do them in a line. Shuffling is OK. So, do this ..
harmonic = 3 //for example. TRY DIFFERENT VALUES
function rh = Random.Range(1,harmonic) (that's 1 not 0)
function rhPosNeg
n = rh
n = either +n or -n
return n
function onePerturbation
{
allBlocks = allBlocks.OrderBy(r => Random.value) //essential
foreach b in allBlocks
newPotentialPosition = Vector2(rhPosNeg,rhPosNeg)
possible = your function to check if it is possible
to have a block at newPotentialPosition,
however be careful not to check "yourself"
if possible, move block to newPotentialPosition
}
The simplest approach is just run onePerturbation, say, three times. Have a look at it between each run. Also try different values of the harmonic tuning factor.
There are many ways to perturb fields of differently-sized blocks, above is a KISS solution that hopefully looks good for your situation.
Coding note...
How to get sets of unique random numbers.
Just to explain this line of code...
List<int> shuf = Enumerable.Range(1,kChunks).OrderBy(r=>Random.value).ToList();
If you are new to coding: say you want to do this: "get a hundred random numbers, from 1 to million, but with no repeats".
Fortunately, this is a very well known problem with a very simple solution.
The way you get numbers with no repeats, is simply shuffle all the numbers, and then take how many you want off the top.
For example, say you need a random couple of numbers from 1-10 but with no repeats.
So, here's the numbers 1-10 shuffled: 3,8,6,1,2,7,10,9,4,5
Simply take what you need off the front: so, 3, 8, 6 etc.
So to make an example let's say you want twelve numbers, no repeats, from 1 through 75. So the first problem is, you want a List with all the numbers up to 75, but shuffled. In fact you do that like this ..
List<int> shuf = Enumerable.Range(1,75).OrderBy(r=>Random.value).ToList();
So that list is 75 items long. You can check it by saying foreach(int r in shuf) Debug.Log(r);. Next in the example you only want 12 of those numbers. Fortunately there's a List call that does this:
shuf = shuf.Take(12)
So, that's it - you now have 12 numbers, no repeats, all random between 1 and 75. Again you can check with foreach(int r in shuf) Debug.Log(r);
In short, when you want "n" numbers, no repeats, between 1 and Max, all you have to so is this:
List<int> shuf = Enumerable.Range(1,Max).OrderBy(r=>Random.value).ToList();
shuf = shuf.Take(n);
et voilà, you can check the result with foreach(int r in shuf) Debug.Log(r);
I just explain this at length because the question is often asked "how to get random numbers that are unique". This is an "age-old" programming trick and the answer is simply that you shuffle an array of all the integers involved.
Interestingly, if you google this question ("how to get random numbers that are unique") it's one of those rare occasions where google is not much help, because: whenever this question is asked, you get a plethora of keen new programmers (who have not heard the simple trick to do it properly!!) writing out huge long complicated ideas, leading to further confusion and complication.
So that's how you make random numbers with no repeats, fortunately it is trivial.

if (elementFound) continue; will skip out this current loop iteration. You need to wrap the int x=Random..; int y=Random()..; part in a while loop with the condition being while(/* position x/y already occupued*/) { /* generate new valid point */} like this for example:
void generateElement(int ratio, int minScale, int maxScale, GameObject g) {
for (int i = 0; i < ratio * generationDefault; i++) {
GameObject el;
// Randomly generate block size and position
bool elementFound = false;
int size, x, y;
do
{
elementFound = false;
size = Random.Range(minScale, maxScale + 1);
x = Random.Range(0, mapSizex + 1 - size);
y = Random.Range(0, mapSizey + 1 - size);
// Check if there is already an element
for (int j = x; j < x + size; j++)
for (int k = y; k < y + size; k++)
if (map[j][k] != null)
elementFound = true;
} while(elementFound);
el = (GameObject)Instantiate(g, new Vector3(x + (float)size / 2, (float)size / 2, y + (float)size / 2), Quaternion.Euler(0, 0, 0));
el.transform.localScale *= size;
// Create element on map array
for (int j = x; j < x + size; j++)
for (int k = y; k < y + size; k++)
if (map[j][k] == null) {
map[j][k] = el.GetComponent<ObjectInterface>();
}
}
}

You shouldn't be getting that many collisions.
Assuming your blocks were ALL 5 units wide and you're trying to fit them into a grid of 500,500 you would have 100*100 spaces for them at minimum, which gives 10,000 spaces into which to fit 1,000 blocks.
Try playing around with this code:
using System;
using System.Collections.Generic;
namespace ConsoleApplication1
{
class Program
{
static void Main()
{
var result = PlaceNonOverlappingBlocks(1000, 5, 500, 500);
}
static List<Block> PlaceNonOverlappingBlocks(int count, int maxBlockSize, int mapX, int mapY)
{
var map = new bool[mapY, mapX];
var rng = new Random();
var result = new List<Block>(count);
int collisions = 0;
while (count > 0)
{
int size = rng.Next(1, maxBlockSize + 1);
int x = rng.Next(0, mapX - size);
int y = rng.Next(0, mapY - size);
if (fits(map, x, y, size))
{
result.Add(new Block(x, y, size));
addToMap(map, x, y, size);
--count;
}
else
{
if (++collisions> 100000)
throw new InvalidOperationException("Hell has frozen over");
}
}
// This is just for diagnostics, and can be removed.
Console.WriteLine($"There were {collisions} collisions.");
return result;
}
static void addToMap(bool[,] map, int px, int py, int size)
{
for (int x = px; x < px+size; ++x)
for (int y = py; y < py + size; ++y)
map[y, x] = true;
}
static bool fits(bool[,] map, int px, int py, int size)
{
for (int x = px; x < px + size; ++x)
for (int y = py; y < py + size; ++y)
if (map[y, x])
return false;
return true;
}
internal class Block
{
public int X { get; }
public int Y { get; }
public int Size { get; }
public Block(int x, int y, int size)
{
X = x;
Y = y;
Size = size;
}
}
}
}

How to group near coordinates

I'm having my thesis "Multiple Choice Examination Checker" and I'm having a big issue about what to do with my problem. I got a picture image (a bitmap specifically) here it is, so you can see:
This is the image with the detectedbox, I will describe this:
This is an examination paper, 1-50 items. each number has a corresponding box(right side of the number, that serves as a container for the answer)
This pictures is just a sample, the number of detected boxes may vary. My approximation is it contains 150-200 detected boxes.
Each detectedboxes are stored in a List(MCvBOX2D) which holds the detectedboxes' size, center, etc.
I transferred those center coordinates in a new list List(PointF) center;
Each box from the image, may have 3-5 detectedboxes. as you can see there are more than one detectedboxes in each of the boxes from the image.
I sorted all the detectedboxes in ascending order, so I would know which will possibly be the number1, number2, and so on..
Here is some of my code, which contains the sorting of boxes.
List<PointF> center = new List<PointF>();
List<PointF> centernew = new List<PointF>();
foreach (MCvBox2D box in lstRectangles)
{
// this code transfers every center-coordinates of detected boxes
// to a new list which is center
center.Add(new PointF(box.center.X, box.center.Y));
}
// and this one sorts the coordinates in ascending order.
centernew = center.OrderBy(p => p.Y).ThenBy(p => p.X).ToList();
I'm done with the sorting part, now my problem is, since there many detected boxes in every box from the image, I would like to group the sortedlist of center-coordinates, so I could eliminate the other detectedboxes and get only one detectedbox for each number.
I know it's hard to understand so I'll explain more.
Let's say my sortedlist of detectedboxes contains first five center-coordinates which are:
let's say this are the center-coordinates of each of the detectedboxes from first box of the image.
center[0] = [ 45.39, 47.6]
center[1] = [ 65.39, 47.6]
center[2] = [ 45.40, 47.10]
center[3] = [ 65.45, 47.25]
center[4] = [ 46.01, 47.50]
and the 2nd are:
center[5] = [ 200.39, 47.2]
center[6] = [ 45.39, 47.2]
center[7] = [ 45.39, 47.3]
center[8] = [ 45.39, 47.55]
My goal is to organize all the sorted detectedboxes inside the list, I must be able to group all the center-coordinates that have close value with the other center, specifically their Y-coordinates.

var rand = new Random();
var threshold = 1;
var points = new List<PointF>();
for (int i = 0; i < 20; i++)
{
points.Add(new PointF((float) rand.NextDouble()*10, (float) rand.NextDouble()*10));
}
Console.WriteLine(points.Count);
for (int i = 0; i < points.Count(); i++)
{
for (int j = i + 1; j < points.Count(); )
{
var pointHere = points[i];
var pointThere = points[j];
var vectorX = pointThere.X - pointHere.X;
var vectorY = pointThere.Y - pointHere.Y;
var length = Math.Sqrt(Math.Pow(vectorX, 2) + Math.Pow(vectorY, 2));
if (length <= threshold)
{
points.RemoveAt(j);
}
else
{
j += 1;
}
}
}
Console.WriteLine(points.Count);

You can calculate the distance between a given point and any other point in the list. If the distance is less than half the width of a box, you can be pretty sure that it's part of the same box.
double threshold = 3.0; // Make this whatever is appropriate
for (int i = center.Count - 1; i >= 0; --i)
if (center.Any(p => p != center[i] && Distance(center[i], p) < threshold))
center.Remove(center[i]);
And you could use this for your Distance() method:
private double Distance(PointF p1, PointF p2)
{
double deltaX = Math.Abs(p1.X - p2.X);
double deltaY = Math.Abs(p1.Y - p2.Y);
return Math.Sqrt((deltaX * deltaX) + (deltaY * deltaY));
}

You could use Distinct with an custom IEqualityComparer (see MSDN).
As an example, define a class:
class BoxEqualityComparer : IEqualityComparer<MCvBox2D>
{
private static Double Tolerance = 0.01; //set your tolerance here
public Boolean Equals(MCvBox2D b1, MCvBox2D b2)
{
if (CentersAreCloseEnough(b1.Center, b2.Center))
{
return true;
}
else
{
return false;
}
}
private Boolean CentersAreCloseEnough(PointF c1, PointF c2)
{
return Math.Abs(c1.X - c2.X) < Tolerance && Math.Abs(c1.Y - c2.Y) < Tolerance;
}
}
then use the method in your code like so:
var distinctRectangles = lstRectangles.Distinct(new BoxEqualityComparer());
You are free to implement CentersAreCloseEnough(PointF c1, PointF c2) however you would like; you could use vector distance, absolute distance in x and y, etc.

If you are just concerned about positions with Y coordinates, just sort by that number. If you want to sort both you can add both X and Y and use that number to sort them. Here is an example of what I meen.
for(int i = 0; i < points.length - 1; i++)
{
int temp = points[i].x + points[i].y;
for(int j = i+1; j < points.length; i++)
{
int temp2 = point[j].x + points[j].y;
if(temp2 < temp)
{
Point jugglePoint = points[i];
points[i] = points[j];
points[j] = jugglePoint;
}
}
}

Variable declared but doesn't show up during Debug C#

screenshot of debug: http://img1.uploadscreenshot.com/images/orig/12/36121481470-orig.jpg
note how x, y have values (i have no idea why x and y stopped on 69 in the for loop - x should've went up to 86 and y to 183) yet r has no value at all. (the variable doesn't exist? what?) how should I fix this?
code if you want to read:
public float[] cht(int[,] matrix)
{
float[] xyrd = new float[4];
int xthreshold, ythreshold;
float slope;
double dir;
float zone;
int[] limitsStorage = new int[3] { matrix.GetLength(0), matrix.GetLength(1), matrix.GetLength(0) / 2 - 10 };
short[,,] accumulator = new short[limitsStorage[0]+1, limitsStorage[1]+1,limitsStorage[2]+1];
for (int x = 0; x < limitsStorage[0]; x++)
{ //first dimension loop of matrix 100
for (int y = 0; y < limitsStorage[1]; y++)
{ //second dimension loop of matrix 120
if (matrix[x, y] == 225)
{//the data at the for loop location is a 1 and not 0 hit.
xthreshold = x - limitsStorage[0] / 2;
ythreshold = y - limitsStorage[1] / 2;
//forget angle, search slope: float angle = xthreshold > 0 ? ((float)Math.Atan2(xthreshold, ythreshold)) : ((float)Math.Atan2(xthreshold, ythreshold) + 180);
slope = xthreshold / ythreshold;
//initiate if loops.
dir = 180 + Math.Round(Math.Atan2(ythreshold, xthreshold) * 57.2957 / 45, 0) * 45 + 45 * (Math.Round(((Math.Atan2(ythreshold, xthreshold) * 57.2957) % 45) / 45));
if (slope > .404 || slope < -.404)
{
if (slope < 2.3558 || slope > -2.3558)
{
if (xthreshold > 0)
{
if (ythreshold > 0)
{
//+x+y zone
zone = 45 + 180;
}
else
{
//+x-y zone
zone = 180 - 45;
}
}
else
{
if (ythreshold > 0)
{
//-x+y zone
zone = 360 - 45;
}
else
{
//-x-y zone
zone = 45;
}
}
}
else if (ythreshold > 0)
{
//+y zone
zone = 360 - 90;
}
else
{
//-y zone
zone = 90;
}
}
else if (xthreshold > 0)
{
//+x zone
zone = 180;
}
else
{
//-x zone
zone = 0;
}
for (int R = 6; R < limitsStorage[2]; R++)
{ //Radius loop for scan 44
float delta = (float)((1 / R) * 57.2957);
for (float Theta = zone - 25; Theta < zone + 25; Theta += delta)
{
accumulator[(int)(((R * Math.Cos(Theta / 57.2957)) < 0 || (R * Math.Cos(Theta / 57.2957)) > limitsStorage[0]) ? 0 : R * Math.Cos(Theta / 57.2957)), (int)(((R * Math.Sin(Theta / 57.2957)) < 0 || (R * Math.Sin(Theta / 57.2957)) > limitsStorage[1]) ? 0 : R * Math.Sin(Theta / 57.2957)),R]++;
//btw, 0,0,R is always the non-center area.
}
}
}
}
}
for (int x = 1; x < limitsStorage[0]; x++)
{
for (int y = 1; y < limitsStorage[1]; y++)
{
for (int r = 6; r < limitsStorage[2]; r++)
{
if (xyrd[3] > accumulator[x, y, r])
{
xyrd[0] = x;
xyrd[1] = y;
xyrd[2] = r;
xyrd[3] = accumulator[x, y, r];
}
}
}
}
if (accPrint)
{
//do something for debugging?
accPrint = false;
}
return xyrd;
}

I just noticed that the x and y have the little lock symbol under them indicating that you have private variables named x and y in the class in which this method is executing. Those are the x and y that you are seeing in the debugger.
r is appropriately out of scope as you've exited the loop in which it is declared.
By the way, x and y are ridiculously bad member variable names, and are ridiculously bad names for for loop variables of type int, especially if they are contained in a class with member variables named x and y.

The only place you declare r is in the for statement, right? That means r goes out of scope as soon as the loop ends. So naturally if you inspect variables at he end of the function, r won't be there.
Confessing I don't know why x and y are in scope based on the comments. They could be class variables, but the asker says not. That's the only explanation I can think of, though.

The behaviour is not weird -- you actually get exactly what you expect.
Please note that the watch window can only accurately show you values that are in scope at the breakpoint.
At the highlighted breakpoint, only accumulator[x, y, r] is in scope, and you see exactly the values you expected.