codeforces 650D D. Image Preview (暴力+二分+dp)

本文解析了CodeForces上的一道题目D.ImagePreview,介绍了题目的背景、要求及限制条件,并提供了详细的解题思路与AC代码实现。

摘要生成于 C知道 ,由 DeepSeek-R1 满血版支持, 前往体验 >

题目链接:
D. Image Preview
time limit per test
1 second
memory limit per test
256 megabytes
input
standard input
output
standard output

Vasya's telephone contains n photos. Photo number 1 is currently opened on the phone. It is allowed to move left and right to the adjacent photo by swiping finger over the screen. If you swipe left from the first photo, you reach photo n. Similarly, by swiping right from the last photo you reach photo 1. It takes a seconds to swipe from photo to adjacent.

For each photo it is known which orientation is intended for it — horizontal or vertical. Phone is in the vertical orientation and can't be rotated. It takes b second to change orientation of the photo.

Vasya has T seconds to watch photos. He want to watch as many photos as possible. If Vasya opens the photo for the first time, he spends 1 second to notice all details in it. If photo is in the wrong orientation, he spends b seconds on rotating it before watching it. If Vasya has already opened the photo, he just skips it (so he doesn't spend any time for watching it or for changing its orientation). It is not allowed to skip unseen photos.

Help Vasya find the maximum number of photos he is able to watch during T seconds.

Input

The first line of the input contains 4 integers n, a, b, T (1 ≤ n ≤ 5·105, 1 ≤ a, b ≤ 1000, 1 ≤ T ≤ 109) — the number of photos, time to move from a photo to adjacent, time to change orientation of a photo and time Vasya can spend for watching photo.

Second line of the input contains a string of length n containing symbols 'w' and 'h'.

If the i-th position of a string contains 'w', then the photo i should be seen in the horizontal orientation.

If the i-th position of a string contains 'h', then the photo i should be seen in vertical orientation.

Output

Output the only integer, the maximum number of photos Vasya is able to watch during those T seconds.

Examples
input
4 2 3 10
wwhw
output
2
input
5 2 4 13
hhwhh
output
4
input
5 2 4 1000
hhwhh
output
5
input
3 1 100 10
whw
output
0
Note

In the first sample test you can rotate the first photo (3 seconds), watch the first photo (1 seconds), move left (2 second), rotate fourth photo (3 seconds), watch fourth photo (1 second). The whole process takes exactly 10 seconds.

Note that in the last sample test the time is not enough even to watch the first photo, also you can't skip it.

题意:给你一个字符串,跟看这张照片花多少时间有关,s[i]==w,则需要1+b,s[i]==h需要1,滑到相邻的照片需要时间a,注意要从第一张开始看,滑到没看的照片时一定要花时间看,滑到已经看过的照片则不需要再花时间看了,由于这个我没读懂而且我一开始还认为不能网回滑,整的我想了一天一夜没想到解决办法,我想过贪心,线段树,单调等等,最后发现真正的题意直接想大哭一场啊啊啊啊啊啊啊啊啊;

思路:一开始有两个方向,向左或者向右,暴力枚举向左走了多少二分查找回来向右能走多少,再暴力枚举一开始就向右走,二分查找能想左走的,这里面最大的就是结果了;

AC代码:

#include <bits/stdc++.h>
using namespace std;
const int N=5e+6;
int n,a,b,t,dp[N],c[N];
char s[N];
int main()
{
    scanf("%d%d%d%d",&n,&a,&b,&t);
    scanf("%s",s+1);
    if(s[1]=='w')t-=b+1;
    else t-=1;
    dp[0]=0;
    for(int i=2;i<=n;i++)
    {
        if(s[i]=='w')dp[i-1]=dp[i-2]+1+b+a;
        else dp[i-1]=dp[i-2]+1+a;
    }
    c[0]=0;
    for(int i=n;i>1;i--)
    {
        if(s[i]=='w')c[n-i+1]=1+b+c[n-i]+a;
        else c[n-i+1]=1+c[n-i]+a;
    }
    if(t<0)
    {
        cout<<"0";
        return 0;
    }
    else {
    int ans=1;
    for(int i=0;i<n;i++)
    {
       if(t>=c[i])
       {
           if(t>c[i]+i*a)
           {
               int pos=lower_bound(dp,dp+n-i,t-c[i]-i*a+1)-dp;//注意是t-c[i]-i*a+1,,不能忘了+1,后面也是;
               ans=max(ans,i+pos);
           }
           else
           {
               ans=max(ans,i+1);
           }
       }
       else break;
    }
    //cout<<ans<<"@"<<endl;
    for(int i=0;i<n;i++)
    {
        if(t>=dp[i])
        {
            if(t>dp[i]+i*a)
            {
                int pos=lower_bound(c,c+n-i,t-dp[i]-i*a+1)-c;
                ans=max(ans,i+pos);
            }
            else ans=max(ans,i+1);
        }
        else break;
    }
    printf("%d\n",ans);
    return 0;
    }
}

 

转载于:https://www.cnblogs.com/zhangchengc919/p/5255355.html

内容概要:该论文聚焦于T2WI核磁共振图像超分辨率问题,提出了一种利用T1WI模态作为辅助信息的跨模态解决方案。其主要贡献包括:提出基于高频信息约束的网络框架,通过主干特征提取分支和高频结构先验建模分支结合Transformer模块和注意力机制有效重建高频细节;设计渐进式特征匹配融合框架,采用多阶段相似特征匹配算法提高匹配鲁棒性;引入模型量化技术降低推理资源需求。实验结果表明,该方法不仅提高了超分辨率性能,还保持了图像质量。 适合人群:从事医学图像处理、计算机视觉领域的研究人员和工程师,尤其是对核磁共振图像超分辨率感兴趣的学者和技术开发者。 使用场景及目标:①适用于需要提升T2WI核磁共振图像分辨率的应用场景;②目标是通过跨模态信息融合提高图像质量,解决传统单模态方法难以克服的高频细节丢失问题;③为临床诊断提供更高质量的影像资料,帮助医生更准确地识别病灶。 其他说明:论文不仅提供了详细的网络架构设计与实现代码,还深入探讨了跨模态噪声的本质、高频信息约束的实现方式以及渐进式特征匹配的具体过程。此外,作者还对模型进行了量化处理,使得该方法可以在资源受限环境下高效运行。阅读时应重点关注论文中提到的技术创新点及其背后的原理,理解如何通过跨模态信息融合提升图像重建效果。
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值