矩形相关思路总结
2020-07-06 本文已影响0人
madao756
leetcode 关于矩形相关的题目真的太多了,把这些思路总结一下
0X00 面积类
最大正方形面积
dp[i][j]
表示以 (i, j) 结尾的最大正方形边长,最后取最大的边长的平方:
class Solution:
def maximalSquare(self, mat: List[List[int]]) -> int:
if not len(mat): return 0
m, n = len(mat), len(mat[0])
dp = [[0] * (1+n) for _ in range(1+m)]
ans = 0
for i in range(1, 1+m):
for j in range(1, 1+n):
v = mat[i-1][j-1]
if v == "0": continue
dp[i][j] = min(dp[i-1][j], dp[i][j-1], dp[i-1][j-1]) + 1
ans = max(ans, dp[i][j])
return ans * ans
最大矩形面积
思路单调栈,需要牢记
class Solution:
def maximalRectangle(self, matrix: List[List[str]]) -> int:
if not len(matrix): return 0
m, n = len(matrix), len(matrix[0])
def helper(hs):
s, res = [-1], 0
for i in range(n):
h = hs[i]
while s[-1] != -1 and hs[s[-1]] >= h:
th = hs[s.pop()]
res = max(res, th * (i - s[-1] - 1))
s.append(i)
while s[-1] != -1:
th = hs[s.pop()]
res = max(res, th * (n - s[-1] - 1))
return res
hs = [0 if matrix[0][i] == "0" else 1 for i in range(n)]
s = helper(hs)
for i in range(1, m):
for j in range(n):
if matrix[i][j] == "0": hs[j] = 0
else: hs[j] += 1
s = max(s, helper(hs))
return s
0X01 个数类
统计正方形个数
与上面 dp 做法一模一样,原因是:最大边长是 n 就会有 n 个正方形,所以把所有相加就是最后的答案
class Solution:
def countSquares(self, mat: List[List[int]]) -> int:
if not len(mat): return 0
m, n = len(mat), len(mat[0])
dp = [[0] * (1+n) for _ in range(1+m)]
ans = 0
for i in range(1, 1+m):
for j in range(1, 1+n):
v = mat[i-1][j-1]
if v == 0: continue
dp[i][j] = min(dp[i-1][j], dp[i][j-1], dp[i-1][j-1]) + 1
ans += dp[i][j]
return ans
统计矩形个数
1504. 统计全 1 子矩形 的做法,这个做法,很好理解但是很难想到:
可以参考这个题解:https://leetcode-cn.com/problems/count-submatrices-with-all-ones/solution/bian-tong-ji-bian-ya-suo-xiang-xi-by-quantum-10/ 写得很好。
思路的关键在于:固定高度,遍历以这个高度出现的矩形的个数
。
class Solution:
def numSubmat(self, mat: List[List[int]]) -> int:
m, n = len(mat), len(mat[0])
ans = 0
for i in range(m):
for j in range(i, m):
now = 0
for k in range(n):
if mat[j][k] == 0: now = 0
else: now = 1 if k == 0 or mat[j][k-1] == 0 else now + 1
ans += now
for j in range(m-1, i, -1):
for k in range(n):
mat[j][k] &= mat[j-1][k]
return ans