<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[weixin_45798949的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/weixin_45798949</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; weixin_45798949]]></copyright><item><title><![CDATA[MDPI模板报错的问题---提示缺少sty文件，并提供bib引入参考文献的方式]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/133035801</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/133035801</guid><author>weixin_45798949</author><pubDate>Tue, 19 Sep 2023 16:54:18 +0800</pubDate><description><![CDATA[平时大多数提交IEEE trans模板时大多使用CTEX编译，然而，MDPI模板需要用texlive，二者之间如果先安装CTEX后安装texlive将会导致库文件的冲突。结果将会报缺少sty的文件错。网上提供了很多解决方案，然而，最有效的则是将CTEX的文件安装目录改个名字。此时再用texworks打开即可编译成功。当需要重新提交IEEE模板时再将ctex文件名改回即可。MDPI模板报错的问题—提示缺少sty文件。]]></description><category></category></item><item><title><![CDATA[YOLOv5训练大规模的遥感实例分割数据集 iSAID从切图到数据集制作及训练]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/129443725</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/129443725</guid><author>weixin_45798949</author><pubDate>Fri, 10 Mar 2023 16:30:06 +0800</pubDate><description><![CDATA[iSAID切图及实例分割训练]]></description><category></category></item><item><title><![CDATA[更改COCO数据集评估标准输出结果，可以自行输入IoU阈值判断最终AP]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/127752343</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/127752343</guid><author>weixin_45798949</author><pubDate>Tue, 08 Nov 2022 16:05:19 +0800</pubDate><description><![CDATA[不同IoU阈值下的COCO评估]]></description><category></category></item><item><title><![CDATA[pytorch中利用图像处理库transform将图像经过变换转换成张量，并可视化处理后的图像]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/118719701</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/118719701</guid><author>weixin_45798949</author><pubDate>Wed, 14 Jul 2021 09:27:23 +0800</pubDate><description><![CDATA[# -*- coding: utf-8 -*-
# -----------------------------------------------------
# Time :  2021/7/14 上午8:39
# Auth :  Written by zuofengyuan
# File :  transform_.py
# Copyright (c) Shenyang Pedlin Technolofy Co., Ltd.
# -------------------------------------]]></description><category></category></item><item><title><![CDATA[pycharm左侧项目文件出现黄色阴影解决]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/114732663</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/114732663</guid><author>weixin_45798949</author><pubDate>Sat, 13 Mar 2021 10:33:58 +0800</pubDate><description><![CDATA[If by “shadow” you mean brown coloring, that means that directory is marked as “excluded” in project structure. The code will run, but code inspection and insight won’t work for excluded folders, so please remove “excluded” flag from Project Structure sett]]></description><category></category></item><item><title><![CDATA[解决ubuntu文件夹带锁问题]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/113735929</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/113735929</guid><author>weixin_45798949</author><pubDate>Sun, 07 Feb 2021 09:56:01 +0800</pubDate><description><![CDATA[cd dir
sudo chown -R username filename

第一步cd到文件夹路径
第二步指定该文件夹带锁文件名
username为电脑使用者名，filename为待解锁文件名
输入后，需要输入密码，即可完成解锁

]]></description><category></category></item><item><title><![CDATA[python修改labelme标注生成的json路径]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/113356366</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/113356366</guid><author>weixin_45798949</author><pubDate>Thu, 28 Jan 2021 19:20:40 +0800</pubDate><description><![CDATA[import json
a=json.load(open(r'E:\pythonpycharm\json----\000001.json'))
print(a['imagePath'])
a['imagePath']=a['imagePath'][a['imagePath'].rfind('\\')+1:]
print(a['imagePath'])


]]></description><category></category></item><item><title><![CDATA[Latex中引用文献，正文里有序号不是乱码问号，文献里无内容]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/110136760</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/110136760</guid><author>weixin_45798949</author><pubDate>Wed, 25 Nov 2020 17:13:11 +0800</pubDate><description><![CDATA[解决：
bib里的文献不要在和文献同一行使用%%注释
在上一行使用即可解决：
如下图正确

下图不正确


]]></description><category></category></item><item><title><![CDATA[python利用try_except打印错误信息，并继续执行]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109635707</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109635707</guid><author>weixin_45798949</author><pubDate>Thu, 12 Nov 2020 08:26:08 +0800</pubDate><description><![CDATA[try:#jia
        print('.............')
    except Exception as e:
        #print('错误类型是',e.__class__.__name__)#except:#jia
        print('错误明细是',e.__class__.__name__,e)   #continue#jia
        continue


]]></description><category></category></item><item><title><![CDATA[latex三线表绘制，合并行与列]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109629057</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109629057</guid><author>weixin_45798949</author><pubDate>Wed, 11 Nov 2020 19:31:00 +0800</pubDate><description><![CDATA[导入包
在最前面导入包(别在正文后面导入，在签名声明)
\usepackage{booktabs}
\usepackage{multirow}

创建一个简单的4行4列表格
%%创建4行4列表格
\begin{table}
\centering
\begin{tabular}{cccc}		
\toprule
x&amp;y&amp;u&amp;8\\
\midrule
5 &amp; 6&amp; z &amp; w \\
8 &amp; 9 &amp; 3&amp; 4 \\
10 &amp; 7 &a]]></description><category></category></item><item><title><![CDATA[将图片补成正方形，不够用黑色补]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109481223</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109481223</guid><author>weixin_45798949</author><pubDate>Tue, 03 Nov 2020 22:45:45 +0800</pubDate><description><![CDATA[import cv2
import os
import numpy as np
def get_resize_square(input,output):#,output
    for filename in os.listdir(input):
        image_path=input+'/'+filename
        image=cv2.imread(image_path)
        print(image)
        x=image.shape[1]##宽度
       ]]></description><category></category></item><item><title><![CDATA[计算图像感兴趣区域(ROI)的像素均值与标准差]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109353424</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109353424</guid><author>weixin_45798949</author><pubDate>Thu, 29 Oct 2020 10:40:41 +0800</pubDate><description><![CDATA[import cv2
import os
#img = cv2.imread(img_path)
#print(img)
#m, s = cv2.meanStdDev(img)
#print('均值',m)
#print('标准差',s)
def read_txt(wei_zhi):###阅读文本文件
    with open(wei_zhi, 'r') as f:
        lines = f.readlines()
    splitlines = [x.strip().split(' ') f]]></description><category></category></item><item><title><![CDATA[使用pytorch自己搭建LetNnet网络并训练一个小的分类器]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109340917</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109340917</guid><author>weixin_45798949</author><pubDate>Wed, 28 Oct 2020 20:18:07 +0800</pubDate><description><![CDATA[1.网络介绍

1、C1层（卷积层）：6@28×28
该层使用了6个卷积核，每个卷积核的大小为5×5，步长为1，这样就得到了6个feature map（特征图），即深度为6。（32-5+1）/1
特征图大小：
每个卷积核（5×5）与原始的输入图像（32×32）进行卷积，这样得到的feature map（特征图）大小为（32-5+1）×（32-5+1）= 28×28
2. 池化层：池化层不改变通道数，
卷积尺寸22，步长为2，通道数不变。输出尺寸：14146
3. 卷积层：
该层使用了16个卷积核，每个卷积核]]></description><category></category></item><item><title><![CDATA[神经网络典型结构Alexnet,VGG,GoogLeNet,Resnet论文]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109322078</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109322078</guid><author>weixin_45798949</author><pubDate>Tue, 27 Oct 2020 22:25:39 +0800</pubDate><description><![CDATA[1.AlexNet网络
Deep Residual Learning for Image Recognition
2.VGG
Very Deep Convolutional Networks for Large-Scale Image Recognition
3.GoogLeNet
Going deeper with convolutions
4.ResNet ：
Deep Residual Learning for Image Recognition

]]></description><category></category></item><item><title><![CDATA[python将mat文件保存成图片，并比较两者数组关系]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109290557</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109290557</guid><author>weixin_45798949</author><pubDate>Mon, 26 Oct 2020 15:22:06 +0800</pubDate><description><![CDATA[import scipy.io as io
import cv2
mat_dir=r'D:\research\ce_Faster-RCNN-TensorFlow-Python3-master\data\VOCdevkit2007\VOC2007\mat-JPEGImaes'#mat路径
a = io.loadmat(mat_dir+'/'+'000001.mat') #加载文件
print('a is:',a['data'])#打印mat文件字典中的值
print('a.shape',a['data'].s]]></description><category></category></item><item><title><![CDATA[python中enumrate,与采用len(list) or循环比较]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109015329</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109015329</guid><author>weixin_45798949</author><pubDate>Sun, 11 Oct 2020 18:57:47 +0800</pubDate><description><![CDATA[seq = ['one', 'two', 'three']
for i, element in enumerate(seq):
    print(i, element)

for i in range(len(seq)):
    print(i,seq[i])

0 one
1 two
2 three

0 one
1 two
2 three


推荐使用enumrate

]]></description><category></category></item><item><title><![CDATA[读txt文档到列表与写列表中元素到txt文档]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109013590</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109013590</guid><author>weixin_45798949</author><pubDate>Sun, 11 Oct 2020 16:39:35 +0800</pubDate><description><![CDATA[根据自己任务进行改写
def read_txt(wei_zhi):
    with open(wei_zhi, 'r') as f:
        lines = f.readlines()
    splitlines = [x.strip().split(' ') for x in lines]
    '''
    T=[]
    for i in range(len(splitlines)):
        sp=splitlines[i]+['qx']
        T.append(]]></description><category></category></item><item><title><![CDATA[2020cvpr最佳人脸识别-Suppressing Uncertainties for Large-Scale Facial Expression Recognition]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109012644</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109012644</guid><author>weixin_45798949</author><pubDate>Sun, 11 Oct 2020 16:34:01 +0800</pubDate><description><![CDATA[原文链接
代码即将开源
Abstact
由于表情模糊、低质量的人脸图像以及注释者的主观性等因素的影响，对一个定性的大规模人脸表情数据集进行注释是非常困难的。这些不确定性导致了深度学习时代大规模人脸表情识别的一个关键挑战。为了解决这一问题，本文提出了一种简单而有效的自修复网络（SCN），它可以有效地抑制不确定性，防止deep网络过度拟合不确定的人脸图像。具体而言，SCN从两个不同的方面抑制了不确定性：1）基于小批量的自我注意机制，对每个训练样本进行排序正则化加权；2）通过仔细的重标记机制，在排名最低的组中修改]]></description><category></category></item><item><title><![CDATA[UnboundLocalError: local variable ‘list_max_score‘ referenced before assignment问题解决]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109012521</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109012521</guid><author>weixin_45798949</author><pubDate>Sun, 11 Oct 2020 15:38:06 +0800</pubDate><description><![CDATA[问题原因：没有提前声明全局变量：
问题解决：找到该变量所在函数位置
在函数下面声明全局变量：
global list_max_score


]]></description><category></category></item><item><title><![CDATA[列表嵌套列表时，输出比较里层列表某个元素的最大值所在的列表]]></title><link>https://blog.youkuaiyun.com/weixin_45798949/article/details/109008212</link><guid>https://blog.youkuaiyun.com/weixin_45798949/article/details/109008212</guid><author>weixin_45798949</author><pubDate>Sun, 11 Oct 2020 09:55:41 +0800</pubDate><description><![CDATA[#列表嵌套列表时，输出比较里层列表某个元素的最大值所在的列表
T=[[ '1', '321.3', '287.1', '352.0', '298.4'], ['0.99', '319.4', '284.5', '354.2', '298.9'], ['1.2', '319.4', '287.2', '351.1', '297.1']]
#目的输出：['0.989', '319.4', '284.5', '354.2', '298.9']
def judege_max_list(list):
    max ]]></description><category></category></item></channel></rss>