摘要
深度补全旨在从激光雷达扫描深度输入的图像中预测物体与相机之间的距离,并将距离表示为密集深度图。扫描深度输入越密集,预测效果越好,但相应的激光雷达设备成本越昂贵,且密集深度输入训练模型在深度输入稀疏时表现较差。同时,训练深度补全模型很难得到密集的准确值。为提高稀疏深度输入模型的性能,提出了一种无监督域自适应方法,对卷积神经网络所生成特征的二阶统计量进行对齐,密集和稀疏的深度输入共享这些特征。
Depth completion aims to predict the distance between objects on an image and the camera capturing the image from a LiDAR scans depth input,and the distance is expressed as a dense depth map.Denser scans depth input leads to better prediction,while the cost of the corresponding LiDAR equipment will be more expensive,and the model trained by dense depth input performs badly on sparse depth input.Meanwhile,it is difficult to get dense ground truth annotations for training depth completion models.To improve the performance of sparse depth input model,an unsupervised domain adaptive method is proposed.The approach aligns the second-order statistics of the features generated by the convolution neural network,which is shared by dense and sparse depth input.
作者
张睿杰
Zhang Ruijie(School of Mechantronics and Vehicle Engineering,Chongqing Jiaotong University,Chongqing 400074,China)
出处
《现代计算机》
2024年第5期77-80,共4页
Modern Computer
关键词
深度自适应
无监督学习
域适应
depth adaptation
unsupervised learning
domain adaptation
