Maximize the binary matrix by filpping submatrix once

Given a binary matrix of R rows and C columns. We are allowed flip to any size of sub matrix. Flipping means changing 1 to 0 and 0 to 1. The task is maximize the number of 1s in the matrix. Output the maximum number of 1s.

Examples:

Input : R = 3, C =3
mat[][] = { 0, 0, 1,
            0, 0, 1,
            1, 0, 1 }

Output : 8
Flip
0 0 1
0 0 1
1 0 1

to get

1 1 1
1 1 1
0 1 1

Input : R = 2, C = 3
mat[][] = { 0, 0, 0,
            0, 0, 0 }
Output : 6



Create a matrix ones[][] of R rows and C columns, which precomputes the number of ones in the submatrix from (0, 0) to (i, j) by

// Common elements in ones[i-1][j] and 
// ones[i][j-1] are ones[i-1][j-1]
ones[i][j] = ones[i-1][j] + ones[i][j-1] - 
             ones[i-1][j-1] + (mat[i][j] == 1)

Since, we are allowed to flip sub matrix only once. We iterate over all possible submatrices of all possible sizes for each cell (i, j) to (i + k – 1, i + k – 1). We calculate the total number of ones after the digits are filliped in the chosen submatrix.

Total number of ones in the final matrix after flipping submatrix (i, j) to (i + k – 1) will be Ones in the whole matrix – Ones in the chosen submatrix + Zeroes in the chosen sub matrix. That comes out to be :-
ones[R][C] – cal(i, j, i + k, j + k – 1) + k*k – cal(i, j, i + k – 1, j + k – 1)
where cal(a, b, c, d) denotes the number of ones in square submatrix of length c – a.

Now cal(x1, y1, x2, y2) can be define by:
ones[x2][y2] – ones[x2][y1 – 1] – ones[x1 – 1][y2] + ones[x1 – 1][y1 – 1].

Below is the C++ implementation of this approach:

// C++ program to find maximum number of ones after
// one flipping in Binary Matrix
#include <bits/stdc++.h>
#define R 3
#define C 3
using namespace std;

// Return number of ones in square submatrix of size
// k x k starting from (x, y)
int cal(int ones[R + 1][C + 1], int x, int y, int k)
{
    return ones[x + k - 1][y + k - 1] - ones[x - 1][y + k - 1]
           - ones[x + k - 1][y - 1] + ones[x - 1][y - 1];
}

// Return maximum number of 1s after flipping a submatrix
int sol(int mat[R][C])
{
    int ans = 0;

    // Precomputing the number of 1s
    int ones[R + 1][C + 1] = {0};
    for (int i = 1; i <= R; i++)
        for (int j = 1; j <= C; j++)
            ones[i][j] = ones[i - 1][j] + ones[i][j - 1] -
                         ones[i - 1][j - 1] +
                         (mat[i - 1][j - 1] == 1);

    // Finding the maximum number of 1s after flipping
    for (int k = 1; k <= min(R, C); k++)
        for (int i = 1; i + k - 1 <= R; i++)
            for (int j = 1; j + k - 1 <= C; j++)
                ans = max(ans, (ones[R][C] + k * k -
                                2 * cal(ones, i, j, k)));
    return ans;
}

// Driver code
int main()
{
    int mat[R][C] = {{0, 0, 1},
        { 0, 0, 1},
        { 1, 0, 1 }
    };

    cout << sol(mat) << endl;

    return 0;
}

Output:

7

Time Complexity: O(R*C*min(R, C))

This article is contributed by Anuj Chauhan. If you like GeeksforGeeks and would like to contribute, you can also write an article using contribute.geeksforgeeks.org or mail your article to contribute@geeksforgeeks.org. See your article appearing on the GeeksforGeeks main page and help other Geeks.

Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above.



My Personal Notes arrow_drop_up



Recommended Posts:



4.6 Average Difficulty : 4.6/5.0
Based on 3 vote(s)






User Actions