傅里叶变换通道注意力网络的胆管癌高光谱图像分割  

Fourier transform channel attention network for cholangiocarcinoma hyperspectral image segmentation

作　　者：郑少佳 邱崧 李庆利[1,2] 周梅[1,2] 胡孟晗[1,2] 于观贞 Zheng Shaojia;Qiu Song;Li Qingli;Zhou Mei;Hu Menghan;Yu Guanzhen(Shanghai Key Laboratory of Multidimensional Information Processing,East China Normal University,Shanghai 200241,China;Engineering Center of SHMEC for Space Information and GNSS,East China Normal University,Shanghai 200241,China)

机构地区：[1]华东师范大学上海市多维度信息处理重点实验室,上海200241 [2]华东师范大学空间信息与定位导航上海高校工程研究中心,上海200241

出　　处：《中国图象图形学报》2021年第8期1836-1846,共11页Journal of Image and Graphics

基　　金：国家自然科学基金项目(61975056);上海市自然科学基金项目(19ZR1416000);上海市国际合作项目(20440713100)。

摘　　要：目的胆管癌高光谱图像的光谱波段丰富但存在冗余,造成基于深度神经网络高光谱图像分割方法的分割精度下降,虽然一些基于通道注意力机制的网络能够关注重要通道,但在处理通道特征时存在信息表示不足问题,因此本文研究构建一种新的通道注意力机制深度网络,以提高分割准确性。方法提出了傅里叶变换多频率通道注意力机制(frequency selecting channel attention,FSCA)。FSCA对输入特征进行2维傅里叶变换,提取部分频率特征,再通过两层全连接层得到通道权重向量,将通道权重与对应通道特征相乘,获得了融合通道注意力信息的输出。针对患癌区域和无癌区域数据不平衡问题引入了Focal损失,结合Inception模块,构建基于Inception-FSCA的胆管癌高光谱图像分割网络。结果在采集的胆管癌高光谱数据集上进行实验,Inception-FSCA网络的准确率(accuracy)、精度(precision)、敏感性(sensitivity)、特异性(specificity)、Kappa系数分别为0.9780、0.9654、0.9586、0.9852、0.9456,优于另外5种对比方法。与合成的假彩色图像的分割结果相比,高光谱图像上的实验指标分别提高了0.0584、0.1058、0.0875、0.0390、0.1493。结论本文所提出的傅里叶变换多频率通道注意力机制能够更有效地利用通道信息,基于Inception-FSCA的胆管癌高光谱图像分割网络能够提升分割效果,在胆管癌医学辅助诊断方面具有研究和应用价值。Objective Cholangiocarcinoma is a rare but highly malignant tumor.Hyperspectral imaging(HSI),which originated from remote sensing,is an emerging image modality for diagnosis and image-guided surgery.HSI takes the advantage of acquiring 2 D images across a wide range of electromagnetic spectrum.HSI can obtain spectral and optical properties of tissue and provide more information than RGB images.Redundant information will persist even though HSI contains tens the amount of data compared with RGB images with the same spatial dimension.Traditional dimensionality reduction methods,such as principal component analysis and kernel method,reduce the data by converting the original spectral space to a low-dimensional one,which is not suitable in end-to-end models.Recently,convolutional neural networks have demonstrated excellent performance on computer vision tasks,including classification,segmentation,and detection.Attention mechanism is used in convolutional neural network(CNN)to improve the representation of feature maps.Typical channel attention modules,such as squeeze-and-excitation net(SENet),squeezes the input features by global average pooling to produce a channel descriptor.However,different channels could have the same mean value.We proposed frequency selecting channel attention(FSCA)mechanism to address this issue.An Inception-FSCA network is also proposed for the segmentation of a hyperspectral image of cholangiocarcinoma tissues.Method FSCA can exploit the information from different frequency components.This method consists of three steps.First,the input feature map is transformed in the frequency domain by Fourier transform.Second,a representative frequency amplitude is selected to efficiently use the obtained frequencies.These selected frequencies are arranged in a column of vectors.Third,these vectors are sent to two consecutive fully connected layers to obtain a channel weight vector.Then,a sigmoid function is used to scale each channel weight between zero and one.Every element in the channel weight vector

关 键 词：胆管癌高光谱图像 卷积神经网络(CNN) 图像分割 通道注意力机制 傅里叶变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]