Finding all adjacent elements in a 2D array

Question

I am working on a project where at one point I am stuck.

My question is for example I have the following 2D array containing 3 different integers.

2 2 2          


        

Answer 1

I love this kind of problems :-) so here it is my answer. As said by Frerich Raabe, this can be solved with a floodFill function. For example, opencv library would provide such a function off the shelf.

Please forgive me if in the following code you'll find traces of C++, in case they should be simple to be replaced.

typedef struct Point {
   int x;
   int y;
} Point;

int areaOfBiggestContiguousRegion(int* mat,int nRows, int nCols) {
  int maxArea = 0;
  int currValue, queueSize, queueIndex;  
  int* aux;
  Point queue[1000]; //Stores the points I need to label
  Point newPoint, currentPoint;
  int x,y,x2,y2;
  //Code: allocate support array aux of same size of mat
  //Code: fill aux of zeros

  for (y = 0; y < nRows; y++)
    for (x = 0; x < nCols; x++)
      if (aux[y * nCols + x] == 0) {//I find a pixel not yet labeled, my seed for the next flood fill
        queueIndex = 0; //Contains the index to the next element in the queue
        queueSize = 0;

        currValue = mat[y * nCols + x]; //The "color" of the current spot
        aux[y * nCols + x] = 1;
        newPoint.x = x;
        newPoint.y = y;
        queue[queueSize] = newPoint;
        queueSize++; 

        while(queueIndex != queueSize) {
          currPoint = queue[queueIndex];
          queueIndex++;

          //Look left, right, up, down

          x2 = currPoint.x - 1;
          y2 = currPoint.y;
          //Some copy & paste, sorry I have been too long on C++ to remember correctly about C functions
          if (x2 >= 0 && aux[y2 * nCols + x2] == 0 && mat[y2 * nCols + x2] == currValue) {
            aux[y2 * nCols + x2] = 1;
            newPoint.x = x2;
            newPoint.y = y2;
            queue[queueSize] = newPoint;
            queueSize++; 
          }

          x2 = currPoint.x + 1;
          y2 = currPoint.y;
          //Some copy & paste, sorry I have been too long on C++ to remember correctly about C functions
          if (x2 < nCols && aux[y2 * nCols + x2] == 0 && mat[y2 * nCols + x2] == currValue) {
            aux[y2 * nCols + x2] = 1;
            newPoint.x = x2;
            newPoint.y = y2;
            queue[queueSize] = newPoint;
            queueSize++; 
          }

          x2 = currPoint.x;
          y2 = currPoint.y - 1;
          //Some copy & paste, sorry I have been too long on C++ to remember correctly about C functions
          if (y2 >= 0 && aux[y2 * nCols + x2] == 0 && mat[y2 * nCols + x2] == currValue) {
            aux[y2 * nCols + x2] = 1;
            newPoint.x = x2;
            newPoint.y = y2;
            queue[queueSize] = newPoint;
            queueSize++; 
          }

          x2 = currPoint.x;
          y2 = currPoint.y + 1;
          //Some copy & paste, sorry I have been too long on C++ to remember correctly about C functions
          if (y2 < nRows && aux[y2 * nCols + x2] == 0 && mat[y2 * nCols + x2] == currValue) {
            aux[y2 * nCols + x2] = 1;
            newPoint.x = x2;
            newPoint.y = y2;
            queue[queueSize] = newPoint;
            queueSize++; 
          }
        } //while

      if (queueSize > maxArea)
        maxArea = queueSize; //If necessary we could store other details like currentValue
      }//if (aux...

return maxArea;
}

Note: In C++ using std containers and a constructor for Point it becomes much more compact