dice_loss.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<mark>主要看这个颜色的内容</mark>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from torch.autograd import Function\n",
    "\n",
    "# class DiceCoeff(Function):\n",
    "#     \"\"\"Dice coeff for individual examples\"\"\"\n",
    "\n",
    "#     def forward(self, input, target):\n",
    "#         #self.save_for_backward(input, target)\n",
    "#         eps = 0.0001\n",
    "#         b = input.size(0)\n",
    "#         self.inter = input.view(b,-1) * target.view(b,-1)\n",
    "#         self.union = torch.sum(input,axis=(1,2,3)) + torch.sum(target,axis=(1,2,3)) + eps\n",
    "\n",
    "#         t = (2 * self.inter.float() + eps) / self.union.float()\n",
    "#         return t.mean(0)\n",
    "\n",
    "class DiceCoeff(Function):\n",
    "    \"\"\"Dice coeff for individual examples\"\"\"\n",
    "\n",
    "    def forward(self, input, target):\n",
    "        self.save_for_backward(input, target)\n",
    "        eps = 0.0001\n",
    "        self.inter = torch.dot(input.view(-1), target.view(-1))\n",
    "        self.union = torch.sum(input) + torch.sum(target) + eps\n",
    "        '''\n",
    "            直接view成一维没问题，因为最后dice coeff是一个scalar，最后还是要在batch中所有sample上求均值。\n",
    "        '''\n",
    "\n",
    "        t = (2 * self.inter.float() + eps) / self.union.float()\n",
    "        return t"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#两个DICE loss的计算结果一样，推荐使用这个！\n",
    "import torch.nn as nn\n",
    "class Dice_loss(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(Dice_loss,self).__init__()\n",
    "\n",
    "    def forward(self,pred,true):\n",
    "        b = pred.size(0)\n",
    "        p = pred.view(b,-1)\n",
    "        t = true.view(b,-1)\n",
    "        coff = 2*(p*t).sum(1)/(p.sum(1)+t.sum(1))\n",
    "        loss = 1- coff.mean(0)\n",
    "        return loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "shape = (4,1,512,512) # bs =4\n",
    "out = torch.randn(shape,dtype=torch.float)\n",
    "probs = torch.sigmoid(out)\n",
    "true_mask = torch.randint(0,2,shape).float()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(tensor([1.]), tensor([-0.0163]), tensor([0.4959]))"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "true_mask[0,...,0,0],out[0,...,0,0],probs[0,...,0,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(0.4998)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dl1 = 1.0 - DiceCoeff().forward(probs,true_mask)\n",
    "dl1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(0.4998)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dl2 = Dice_loss().forward(probs,true_mask)\n",
    "dl2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<mark>由此说明两个DICE Loss的计算结果一样！</mark> <br>\n",
    "<mark>并且这里的计算结果来自于bs=4的tensor</mark> <br>\n",
    "这里的两个DICE loss都是正确的，都能直接用。<br>\n",
    "下面这个也能直接用。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "\n",
    "class Dice_coeff(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.eps = 1e-6\n",
    "\n",
    "    def forward(self, pred, target):\n",
    "        b = pred.size(0)\n",
    "        p = pred.view(b,-1)\n",
    "        t = target.view(b,-1)\n",
    "        coff = 2*(p*t).sum(1)/(p.sum(1)+t.sum(1))\n",
    "#         loss = 1- coff.mean(0)\n",
    "        return coff"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-------------------------------"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "\n",
    "class DICEandfocalLoss(nn.Module):\n",
    "    def __init__(self,device,num_classes=2):\n",
    "        super().__init__()\n",
    "        self.device = device\n",
    "        self.num_classes = num_classes\n",
    "        \n",
    "\n",
    "    def forward(self,probs,true_mask):\n",
    "        #-------------------------multi-class\n",
    "        # 此时的probs = softmax(logits)，因为是多分类，所以用softmax\n",
    "        pred = probs>0.5\n",
    "        classes_points = []\n",
    "        for i in range(self.num_classes):\n",
    "            classes_points.append(pred==i)\n",
    "\n",
    "        \n",
    "        TP = torch.zeros((self.num_classes),dtype=torch.float32).to(self.device)\n",
    "        FP = torch.zeros((self.num_classes),dtype=torch.float32).to(self.device)\n",
    "        FN = torch.zeros((self.num_classes),dtype=torch.float32).to(self.device)\n",
    "        \n",
    "        loss_focal = torch.zeros((self.num_classes,),dtype=torch.float32).to(self.device)\n",
    "\n",
    "        for i in range(self.num_classes):\n",
    "            #下注释代码为旧的错误代码，之前没有很好的理解论文中ΣN 的意思。新代码理解正确了✔\n",
    "            #新理解即：Σn p(c) 是指求和 HW平面上所有pixel不管这个类别c是不是argmax @，即@的预测类)\n",
    "#             TP[i] = ((true_mask*probs)*classes_points[i]).sum(axis=(1,2,3))\n",
    "#             FN[i] = ((true_mask*(1.0-probs))*classes_points[i]).sum(axis=(1,2,3))\n",
    "#             FP[i] = (((1.0-true_mask)*probs)*classes_points[i]).sum(axis=(1,2,3))\n",
    "            TP[i] = (probs[:,i] * true_mask[:,i]).sum(axis=(1,2)).mean()\n",
    "            FN[i] = ((1.0-probs[:,i]) * true_mask[:,i]).sum(axis=(1,2)).mean()\n",
    "            FP[i] = (probs[:,i] * (1-true_mask[:,i])).sum(axis=(1,2)).mean()\n",
    "            #print(FP[i].shape)\n",
    "            \n",
    "            loss_focal += (classes_points[i]*(1-probs[:,i].pow(2))*probs[:,i].log()).mean(axis=(0,2,3))\n",
    "            \n",
    "        \n",
    "        \n",
    "        #print()\n",
    "        \n",
    "        alpha = beta = 0.5\n",
    "        lambaa = 1\n",
    "        \n",
    "        # change here：分子分母都*2 && α=β=1，则结果和链接的多分类DICE Loss一致。\n",
    "        # 分子分母的TP都不乘2 && α=β=0.5，则对应原论文的公式！\n",
    "        #loss_dice = self.num_classes - (2*TP / (2*TP+alpha*FN+beta*FP)).sum()\n",
    "        loss_dice = self.num_classes - (TP / (TP+alpha*FN+beta*FP)).sum()\n",
    "        loss_focal = -loss_focal.sum()\n",
    "        #这里的focal loss仅支持二分类，而DICE Loss则只支持多分类(看最下测试代码易懂)。\n",
    "#         loss1 = (true_mask*(1-probs)**2)*torch.log(probs)\n",
    "#         loss2 = (1-true_mask)*(probs**2)*torch.log(1-probs)\n",
    "#         loss_focal = (loss1+loss2).mean(axis=(1,2,3)) * lambdaa\n",
    "#         loss_focal = -loss_focal.mean()\n",
    "        \n",
    "        loss = loss_dice.mean()+loss_focal\n",
    "        #-------------------------multi-class\n",
    "        \n",
    "        #-------------------------two-class\n",
    "        #参考上面cell中的class Dice_loss的forward函数就行了\n",
    "        #-------------------------multi-class\n",
    "\n",
    "        '''\n",
    "            由https://blog.csdn.net/liangjiu2009/article/details/107352164 关于多分类DICE loss的代码(👇)易懂,\n",
    "            常见的多分类DICE Loss其实就是👇论文公式中，当取α=β=1时，对应的公式。\n",
    "            所以我这个class的参数self.num_classes=num,其实对应了？？ \n",
    "            *所以论文中的公式可以看成常用多分类DICE loss的泛化版，即引入了两个参数α和β来使之适应更多不同的情况。*\n",
    "            *这有点像那篇通过引入新参数统一了GAP和GMP为一个公式的那个方法 nm*\n",
    "            @所以结论是，在我毕业设计的这个问题中，我应该使用two-class的loss，而不应该使用multi-class的loss@\n",
    "            @这也是我计算为什么出错的原因@\n",
    "        '''\n",
    "        '''\n",
    "            # 遍历 channel，得到每个类别的二分类 DiceLoss\n",
    "            for i in range(C):\n",
    "                dice_loss = binaryDiceLoss(logits[:, i], target[:, i])\n",
    "                total_loss += dice_loss.ipynb\n",
    "        '''\n",
    "        #虽然叫做DICEandFocalLoss，但是这里为了比较DICE loss部分是否正确，所以最终只返回其中的loss_dice部分\n",
    "        #focal loss部分已验证正确！见focal_loss.ipynb即可\n",
    "        return loss_dice.mean()"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![image.png](attachment:image.png)\n",
    "AnatomyNet: Deep Learning for Fast and Fully Automated Whole-volume\n",
    "Segmentation of Head and Neck Anatomy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "class BinaryDiceLoss(nn.Module):\n",
    "\tdef __init__(self):\n",
    "\t\tsuper(BinaryDiceLoss, self).__init__()\n",
    "\t\n",
    "\tdef forward(self, input, targets):\n",
    "\t\t# 获取每个批次的大小 N\n",
    "\t\tN = targets.size()[0]\n",
    "\t\t# 平滑变量\n",
    "\t\tsmooth = 1\n",
    "\t\t# 将宽高 reshape 到同一纬度\n",
    "\t\tinput_flat = input.view(N, -1)\n",
    "\t\ttargets_flat = targets.view(N, -1)\n",
    "\t\n",
    "\t\t# 计算交集\n",
    "\t\tintersection = input_flat * targets_flat \n",
    "\t\tN_dice_eff = (2 * intersection.sum(1) + smooth) / (input_flat.sum(1) + targets_flat.sum(1) + smooth)\n",
    "\t\t# 计算一个批次中平均每张图的损失\n",
    "\t\tloss = 1 - N_dice_eff.sum() / N\n",
    "\t\treturn loss\n",
    "\n",
    "    \n",
    "\n",
    "class MultiClassDiceLoss(nn.Module):\n",
    "\tdef __init__(self, weight=None, ignore_index=None, **kwargs):\n",
    "\t\tsuper(MultiClassDiceLoss, self).__init__()\n",
    "\t\tself.weight = weight\n",
    "\t\tself.ignore_index = ignore_index\n",
    "\t\tself.kwargs = kwargs\n",
    "\t\n",
    "\tdef forward(self, input, target):\n",
    "\t\t\"\"\"\n",
    "\t\t\tinput tesor of shape = (N, C, H, W)\n",
    "\t\t\ttarget tensor of shape = (N, H, W)\n",
    "\t\t\"\"\"\n",
    "\t\t# 先将 target 进行 one-hot 处理，转换为 (N, C, H, W)\n",
    "\t\tnclass = input.shape[1]\n",
    "\t\t#target = one_hot(target.long(), nclass)\n",
    "\n",
    "\t\tassert input.shape == target.shape, \"predict & target shape do not match\"\n",
    "\t\t\n",
    "\t\tbinaryDiceLoss = BinaryDiceLoss()\n",
    "\t\ttotal_loss = 0\n",
    "\t\t\n",
    "\t\t# 归一化输出\n",
    "\t\tlogits = F.softmax(input, dim=1)\n",
    "\t\tC = target.shape[1]\n",
    "\t\t\n",
    "\t\t# 遍历 channel，得到每个类别的二分类 DiceLoss\n",
    "\t\tfor i in range(C):\n",
    "\t\t\tdice_loss = binaryDiceLoss(logits[:, i], target[:, i])\n",
    "\t\t\ttotal_loss += dice_loss\n",
    "\t\t\n",
    "\t\t# 每个类别的平均 dice_loss\n",
    "\t\treturn total_loss\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "https://blog.csdn.net/liangjiu2009/article/details/107352164 "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "------------------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<mark>这里的测试表明了我写的DICE Loss也能用到多分类中</mark>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device('cpu')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "out = torch.randn((4,2,512,512),dtype=torch.float)\n",
    "probs = torch.nn.functional.softmax(out,dim=1)\n",
    "true_mask = torch.randint(0,2,(4,2,512,512)).float()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(0.9999)"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "multi_class_dl1 = DICEandfocalLoss(device,2)\n",
    "mcdl1 = multi_class_dl1(probs,true_mask)\n",
    "mcdl1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "链接https://blog.csdn.net/liangjiu2009/article/details/107352164 和论文中的loss是等价的！只需要根据“change here”对论文公式做出一些修改就能使其变为DICE LOSS。<br>\n",
    "这也从另外一个角度说明，论文中的是：C-ΣTP/(TP+FN+FP)并不是DICE LOSS(F1)，而是一种变种！(反正它们做实验发现有效 nm为什么变)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(0.9999)"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "multi_class_dl2 = MultiClassDiceLoss()\n",
    "mcdl2 = multi_class_dl2(out,true_mask)\n",
    "mcdl2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(tensor(0.9999), tensor(0.9999))"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 按照“change here”修改后\n",
    "# 上面已经去除了修改，注意。\n",
    "mcdl1, mcdl2"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:torch]",
   "language": "python",
   "name": "conda-env-torch-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.10"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}