《Gradient Harmonized Single-stage Detector》论文阅读以及代码分析

由于香港中文大学多媒体实验室在深度学习时代发表论文的数量以及质量都是值得称赞的，故此对于算法工程师好好学习下https://github.com/open-mmlab/mmdetection是很有必要的。这篇论文主要是引入了新的loss解决物体检测中positive example和negative example的问题以及easy example和hard example的问题。首先先贴出mmdetection中关于GHM loss的代码：

import torch
import torch.nn as nn
import torch.nn.functional as Ffrom ..registry import LOSSESdef _expand_binary_labels(labels, label_weights, label_channels):bin_labels = labels.new_full((labels.size(0), label_channels), 0)inds = torch.nonzero(labels >= 1).squeeze()if inds.numel() > 0:bin_labels[inds, labels[inds] - 1] = 1bin_label_weights = label_weights.view(-1, 1).expand(label_weights.size(0), label_channels)return bin_labels, bin_label_weights# TODO: code refactoring to make it consistent with other losses
@LOSSES.register_module
class GHMC(nn.Module):"""GHM Classification Loss.Details of the theorem can be viewed in the paper"Gradient Harmonized Single-stage Detector".https://arxiv.org/abs/1811.05181Args:bins (int): Number of the unit regions for distribution calculation.momentum (float): The parameter for moving average.use_sigmoid (bool): Can only be true for BCE based loss now.loss_weight (float): The weight of the total GHM-C loss."""def __init__(self, bins=10, momentum=0, use_sigmoid=True, loss_weight=1.0):super(GHMC, self).__init__()self.bins = binsself.momentum = momentumself.edges = torch.arange(bins + 1).float().cuda() / binsself.edges[-1] += 1e-6if momentum > 0:self.acc_sum = torch.zeros(bins).cuda()self.use_sigmoid = use_sigmoidif not self.use_sigmoid:raise NotImplementedErrorself.loss_weight = loss_weightdef forward(self, pred, target, label_weight, *args, **kwargs):"""Calculate the GHM-C loss.Args:pred (float tensor of size [batch_num, class_num]):The direct prediction of classification fc layer.target (float tensor of size [batch_num, class_num]):Binary class target for each sample.label_weight (float tensor of size [batch_num, class_num]):the value is 1 if the sample is valid and 0 if ignored.Returns:The gradient harmonized loss."""# the target should be binary class labelif pred.dim() != target.dim():target, label_weight = _expand_binary_labels(target, label_weight, pred.size(-1))target, label_weight = target.float(), label_weight.float()edges = self.edgesmmt = self.momentumweights = torch.zeros_like(pred)# gradient lengthg = torch.abs(pred.sigmoid().detach() - target)valid = label_weight > 0tot = max(valid.float().sum().item(), 1.0)n = 0  # n valid binsfor i in range(self.bins):inds = (g >= edges[i]) & (g < edges[i + 1]) & validnum_in_bin = inds.sum().item()if num_in_bin > 0:if mmt > 0:self.acc_sum[i] = mmt * self.acc_sum[i] \+ (1 - mmt) * num_in_binweights[inds] = tot / self.acc_sum[i]else:weights[inds] = tot / num_in_binn += 1if n > 0:weights = weights / nloss = F.binary_cross_entropy_with_logits(pred, target, weights, reduction='sum') / totreturn loss * self.loss_weight# TODO: code refactoring to make it consistent with other losses
@LOSSES.register_module
class GHMR(nn.Module):"""GHM Regression Loss.Details of the theorem can be viewed in the paper"Gradient Harmonized Single-stage Detector"https://arxiv.org/abs/1811.05181Args:mu (float): The parameter for the Authentic Smooth L1 loss.bins (int): Number of the unit regions for distribution calculation.momentum (float): The parameter for moving average.loss_weight (float): The weight of the total GHM-R loss."""def __init__(self, mu=0.02, bins=10, momentum=0, loss_weight=1.0):super(GHMR, self).__init__()self.mu = muself.bins = binsself.edges = torch.arange(bins + 1).float().cuda() / binsself.edges[-1] = 1e3self.momentum = momentumif momentum > 0:self.acc_sum = torch.zeros(bins).cuda()self.loss_weight = loss_weight# TODO: support reduction parameterdef forward(self, pred, target, label_weight, avg_factor=None):"""Calculate the GHM-R loss.Args:pred (float tensor of size [batch_num, 4 (* class_num)]):The prediction of box regression layer. Channel number can be 4or 4 * class_num depending on whether it is class-agnostic.target (float tensor of size [batch_num, 4 (* class_num)]):The target regression values with the same size of pred.label_weight (float tensor of size [batch_num, 4 (* class_num)]):The weight of each sample, 0 if ignored.Returns:The gradient harmonized loss."""mu = self.muedges = self.edgesmmt = self.momentum# ASL1 lossdiff = pred - targetloss = torch.sqrt(diff * diff + mu * mu) - mu# gradient lengthg = torch.abs(diff / torch.sqrt(mu * mu + diff * diff)).detach()weights = torch.zeros_like(g)valid = label_weight > 0tot = max(label_weight.float().sum().item(), 1.0)n = 0  # n: valid binsfor i in range(self.bins):inds = (g >= edges[i]) & (g < edges[i + 1]) & validnum_in_bin = inds.sum().item()if num_in_bin > 0:n += 1if mmt > 0:self.acc_sum[i] = mmt * self.acc_sum[i] \+ (1 - mmt) * num_in_binweights[inds] = tot / self.acc_sum[i]else:weights[inds] = tot / num_in_binif n > 0:weights /= nloss = loss * weightsloss = loss.sum() / totreturn loss * self.loss_weight

to be continue

《Gradient Harmonized Single-stage Detector》论文阅读以及代码分析相关推荐

DSSD : Deconvolutional Single Shot Detector论文阅读笔记
文章目录 DSSD : Deconvolutional Single Shot Detector论文阅读笔记2017 Abstract 1. Introduction 2. Related Work ...
目标检测--Accurate Single Stage Detector Using Recurrent Rolling Convolution
Accurate Single Stage Detector Using Recurrent Rolling Convolution CVPR 2017 商汤科技关于目标检测的文献 Code: htt ...
EAST: An Efﬁcient and Accurate Scene Text Detector 论文阅读
EAST: An Efﬁcient and Accurate Scene Text Detector 论文阅读 Reference 正文摘要引言相关工作方法算法网络设计标签生成损失函数 ...
OpenCV图像处理算法——7（《Contrast image correction method》论文阅读及代码实现）
<Contrast image correction method> 论文阅读及代码实现以下内容大部分引自:https://cloud.tencent.com/developer/art ...
fishnet：论文阅读与代码理解
fishnet:论文阅读与代码理解一.论文概述二.整体框架三.代码理解四.总结 fishnet论文地址:http://papers.nips.cc/paper/7356-fishnet-a-v ...
CSI笔记【5】：Widar2.0: Passive Human Tracking with a Single Wi-Fi Link论文阅读
CSI笔记[5]:Widar2.0: Passive Human Tracking with a Single Wi-Fi Link论文笔记前言 Abstract 1 INTRODUCTION 2 ...
Deep Depth Completion of a Single RGB-D Image论文阅读记录以及quicktest
(一)论文简要说明这是一篇2018年CVPR的最新论文,可以直接通过输入RGB图以及相对应的Depth图,然后可以直接补全任意形式深度图的缺失. 论文地址:https://arxiv.org/abs ...
九月学习笔记（FM、一些论文阅读、代码）
目录 2020.09.16 FM 因子分解机 2021.09.18 论文阅读 Interactive Recommender System via Knowledge Graph-enhanced R ...
paperswithcode 论文阅读与代码复现
Machine Learning论文阅读与复现神奇宝贝 1.丰富的论文合集 2.丰富的数据集 3.方法合集 4.论文解析要是有一个cs科研er不知道这个宝藏网站,我都会伤心的,OK?https:/ ...

《Gradient Harmonized Single-stage Detector》论文阅读以及代码分析

《Gradient Harmonized Single-stage Detector》论文阅读以及代码分析相关推荐

最新文章

热门文章