<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[xuru_0927的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/xuru_0927</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; xuru_0927]]></copyright><item><title><![CDATA[Constrained Probabilistic Mask Learning for Task-specific Undersampled MRI Reconstruction]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/138379202</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/138379202</guid><author>xuru_0927</author><pubDate>Mon, 13 May 2024 21:45:49 +0800</pubDate><description><![CDATA[【代码】Constrained Probabilistic Mask Learning for Task-specific Undersampled MRI Reconstruction。]]></description><category></category></item><item><title><![CDATA[UserWarning: FigureCanvasAgg is non-interactive, and thus cannot be shown]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/133516829</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/133516829</guid><author>xuru_0927</author><pubDate>Tue, 03 Oct 2023 13:12:36 +0800</pubDate><description><![CDATA[【代码】UserWarning: FigureCanvasAgg is non-interactive, and thus cannot be shown。]]></description><category></category></item><item><title><![CDATA[install GitHub Desktop on ubuntu]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/133396229</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/133396229</guid><author>xuru_0927</author><pubDate>Thu, 28 Sep 2023 23:59:01 +0800</pubDate><description><![CDATA[【代码】install GitHub Desktop on ubuntu。]]></description><category></category></item><item><title><![CDATA[UNet笔记]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/123886603</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/123886603</guid><author>xuru_0927</author><pubDate>Thu, 07 Apr 2022 01:33:27 +0800</pubDate><description><![CDATA[import os

task_name = 'Task114_heart_MNMs'
raw_data_base_path="/home/ilkay/Documents/ruru/nnUNet/ocmr_task114/nnUNet_raw"
preprocessed_path="/home/ilkay/Documents/ruru/nnUNet/ocmr_task114/nnUNet_preprocessed"
result_folder_path="/home/ilkay/Documents/ruru]]></description><category></category></item><item><title><![CDATA[进展---一些笔记]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/123432342</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/123432342</guid><author>xuru_0927</author><pubDate>Tue, 15 Mar 2022 14:00:12 +0800</pubDate><description><![CDATA[把OCMR的fs的shape写进csv文件
import numpy as np
import math
from ismrmrdtools import transform
import ocmr_unet.read_ocmr as read
import os

files = sorted(list(pathlib.Path('/home/ilkay/Documents/ruru/pg_mri/datasets/ocmr/train/').iterdir()))
out_list = []
for f]]></description><category></category></item><item><title><![CDATA[SeqMRI train process]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/121187508</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/121187508</guid><author>xuru_0927</author><pubDate>Fri, 12 Nov 2021 19:25:48 +0800</pubDate><description><![CDATA[SeqMRI train processLOUPE大体流程train的过程从初始化mask到得到zero-filled imagereconstruction过程怎么得到masksequential网络
LOUPE大体流程
LOUPE(
  (samplers): ModuleList(
    (0): LOUPESampler(
      (gen_mask): LineConstrainedProbMask()
    )
  )
  (reconstructor): LOUPEUNet(
    ]]></description><category></category></item><item><title><![CDATA[DL-HW]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/120847247</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/120847247</guid><author>xuru_0927</author><pubDate>Tue, 19 Oct 2021 15:25:05 +0800</pubDate><description><![CDATA[DL homeworkHW0linear_regression:
HW0linear_regression:
https://www.cs.toronto.edu/~frossard/post/linear_regression/

]]></description><category></category></item><item><title><![CDATA[c1_greedy_policy_processing]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/120464989</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/120464989</guid><author>xuru_0927</author><pubDate>Fri, 01 Oct 2021 07:49:31 +0800</pubDate><description><![CDATA[def get_policy_probs(model, recons, mask): #mask(3,1,1,256,1) recons(3,1,256,256)
    channel_size = mask.shape[1] #1
    res = mask.size(-2) #256
    recons = recons.view(mask.size(0) * channel_size, 1, res, res) #(3,1,256,256)


    # Obtain policy model]]></description><category></category></item><item><title><![CDATA[可视化代码]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119831881</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119831881</guid><author>xuru_0927</author><pubDate>Mon, 30 Aug 2021 03:48:51 +0800</pubDate><description><![CDATA[可视化代码real转complexcomplex转realdpi
real转complex
def real2complex(x):
    x = x[:,0] + (1j * x[:,1])  
    return x

complex转real
def complex2real(x):
    '''
    Parameter
    ---------
    x: ndarray
        assumes at least 2d. Last 2D axes are split in te]]></description><category></category></item><item><title><![CDATA[pg_mri----OCMR policy network]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119979107</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119979107</guid><author>xuru_0927</author><pubDate>Mon, 30 Aug 2021 03:48:14 +0800</pubDate><description><![CDATA[before policy network作为initial
step0
经过policy network输出变量output是(1,255)，然后排出中间的白条，取其余的最大值位置第145。
因为8个trajectories，这里复制出了8个channel得到(8,256)，每个都是(255,255),
所以8个相同的mask_kspace
step1
这里actions = policy.sample()tensor([[144, 144, 146, 144, 144, 146, 144, 146]].]]></description><category></category></item><item><title><![CDATA[pg_mri----brain policy network]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119933953</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119933953</guid><author>xuru_0927</author><pubDate>Fri, 27 Aug 2021 13:49:02 +0800</pubDate><description><![CDATA[train_policy.pyCreate data loadertest_loader = create_data_loader()test_data_range_dict = create_data_range_dict()do_and_log_evaluation一直往mask中加竖线，for循环16遍step=0step=1step=2,3......15循环完所有batch计算ssim和psnr
Create data loader
test_loader = create_data_loader]]></description><category></category></item><item><title><![CDATA[pg_mri----brain reconstruction]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119778017</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119778017</guid><author>xuru_0927</author><pubDate>Sat, 21 Aug 2021 14:42:19 +0800</pubDate><description><![CDATA[train-reconstructiontrain_loader = create_data_loader(args, 'train', shuffle=True)dev_loader = create_data_loader(args, 'val')display_loader = create_data_loader(args, 'val', display=True)scheduler = torch.optim.lr_scheduler.StepLR(optimizer, args.lr_step_]]></description><category></category></item><item><title><![CDATA[我遇到的疑惑及解惑]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119079243</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119079243</guid><author>xuru_0927</author><pubDate>Tue, 17 Aug 2021 13:43:37 +0800</pubDate><description><![CDATA[我遇到的疑惑及解惑Why Zero-padding值得反复看的视频fft, ifft， fftshift，ifftshift
Why Zero-padding
Zero-padding methods to improve the DFT
值得反复看的视频

MRI Image Parameters and Optimization
Active MR k-space Sampling with Reinforcement Learning https://arxiv.org/abs/2007.10469
]]></description><category></category></item><item><title><![CDATA[cascade MRI reconstruction:evaluate]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119272215</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119272215</guid><author>xuru_0927</author><pubDate>Mon, 02 Aug 2021 21:42:09 +0800</pubDate><description><![CDATA[if draw:
    pred = pred[0]
    grnd = grnd[0]
    err = np.abs(pred - grnd) ** 2
    err -= err.min()
    err /= err.max()
    err *= 255
    err = err.astype(np.uint8)
    err = cv2.applyColorMap(err, 2)
    plt.imsave(os.path.join(kwargs['output_path'],]]></description><category></category></item><item><title><![CDATA[model.load_state_dict(state_dict)报错问题]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119274321</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119274321</guid><author>xuru_0927</author><pubDate>Sat, 31 Jul 2021 13:17:29 +0800</pubDate><description><![CDATA[看一下训练得到的模型参数
    state_dict = torch.load('logs/sanity_3/checkpoint', map_location='cuda' if args['train']['cuda'] else 'cpu')
    state_dict = state_dict['model']

看一下参数
for k,v in state_dict.items():
    print(k)

输出：
module.block.0.layers.0.weight
modul.]]></description><category></category></item><item><title><![CDATA[MSE & PSNR & SSIM]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119145862</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119145862</guid><author>xuru_0927</author><pubDate>Sat, 31 Jul 2021 09:33:58 +0800</pubDate><description><![CDATA[PSNR &amp; SSIMcascade MRI reconstruction: PSNRbase_psnrtest_psnrsummarycascade MRI reconstruction: SSIM

Understanding Regression - Mean Squared Error / Loss
In Part 1 of this video,  calculation of MSE is explained and in the following Part 2 of the vide]]></description><category></category></item><item><title><![CDATA[cascade MRI reconstruction:loss function]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119237215</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119237215</guid><author>xuru_0927</author><pubDate>Fri, 30 Jul 2021 17:12:46 +0800</pubDate><description><![CDATA[MSELOSS大体流程手动计算粗略手动计算调torchstep by step手动计算用numpysummary总结数据集值扩大1000倍
大体流程
用了10个.mat，让train为6个，test为4个，batch_size=2
loss = criterion(output['image'], batch['full'])





执行loss跳到这里

再跳到这里

再跳到这里

再跳到这里


跳到这里


执行上面的return语句时跳到这里

然后return结束返回到原位置


跳回到原来的]]></description><category></category></item><item><title><![CDATA[iTerm2]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119193985</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119193985</guid><author>xuru_0927</author><pubDate>Thu, 29 Jul 2021 12:55:40 +0800</pubDate><description><![CDATA[iTerm2Oh My ZshThemeszsh Auto SuggestionsSyntax HighlightingWord JumpsChange the username@hostname终端运行.ipynb
export https_proxy=http://127.0.0.1:7890 http_proxy=http://127.0.0.1:7890 all_proxy=socks5://127.0.0.1:789

Oh My Zsh
https://ohmyz.sh
Install with]]></description><category></category></item><item><title><![CDATA[2021-07-26科研进展]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119114425</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119114425</guid><author>xuru_0927</author><pubDate>Mon, 26 Jul 2021 17:47:51 +0800</pubDate><description><![CDATA[The words in blue are the knowledge points involved in the presentation
In preprocessing stage：
First, the original data is preprocessed using zero-padding
Then the fully sampled original data is accelerated to obtain the undersampled data
In order to per.]]></description><category></category></item><item><title><![CDATA[Parallel MRI]]></title><link>https://blog.youkuaiyun.com/xuru_0927/article/details/119062698</link><guid>https://blog.youkuaiyun.com/xuru_0927/article/details/119062698</guid><author>xuru_0927</author><pubDate>Sat, 24 Jul 2021 19:30:22 +0800</pubDate><description><![CDATA[Contents4.1 Introduction4.2 Parallel Imaging as an Inverse Problem .4.2.1Forward Model ………4.2.2Image Reconstruction ………4.2.3Regularization……4.2.4Discretization………………4.2.5Numerical Optimization …4.3 Autocalibration ……4.3.1 Nonlinear Inverse Reconstruction .]]></description><category></category></item></channel></rss>