Weakly Supervised Spatial Deep Learning based on Imperfect Vector Labels with Registration Errors
Weakly Supervised Spatial Deep Learning based on Imperfect Vector Labels with Registration Errors
复制标题
DOI：
10.1145/3447548.3467301
复制
发表时间：
2021-08
期刊：
影响因子：
--
通讯作者：
Zhe Jiang;Wenchong He;M. Kirby;S. Asiri;Dan Yan
中科院分区：
文献类型：
--
作者：
Zhe Jiang;Wenchong He;M. Kirby;S. Asiri;Dan Yan
研究了基于不完美向量训练标签的空间栅格数据弱监督学习。给定光栅特征图像和带有位置配准错误的不完美（弱）矢量标签，我们的目标是学习一个深度学习模型来进行像素分类，同时优化矢量标签。这个问题在许多地球科学应用中很重要，例如从地球图像中绘制流线和道路图，其中注释不完美的粗糙矢量标签比绘制精确标签更有效。但由于矢量标签与光栅特征像素的不对齐以及在学习神经网络参数时需要推断真实矢量标签位置，因此该问题具有挑战性。弱监督学习的现有工作通常集中在标签语义中的噪音和错误，假设标签位置是正确的或不相关的（例如，独立分布）。有一些工作存在的标签注册错误，但这些方法往往集中在标签错位的对象段边界上的像素级，而不保证矢量的连续性。为了填补这一差距，本文提出了一种基于期望最大化的空间学习框架，该框架在推断真实矢量标签位置的同时迭代更新深度神经网络参数。具体而言，真实矢量位置的推断基于当前像素类预测和矢量的几何属性。在国家水文数据集（NHD）细化的真实世界的高分辨率遥感数据集的评估表明，该框架优于基线方法的分类精度和细化的矢量质量。
This paper studies weakly supervised learning on spatial raster data based on imperfect vector training labels. Given raster feature imagery and imperfect (weak) vector labels with location registration errors, our goal is to learn a deep learning model for pixel classification and refine vector labels simultaneously. The problem is important in many geoscience applications such as streamline delineation and road mapping from earth imagery, where annotating imperfect coarse vector labels is far more efficient than drawing precise labels. But the problem is challenging due to the misalignment of vector labels with raster feature pixels and the need to infer true vector label location while learning neural network parameters. Existing works on weakly supervised learning often focus on noise and errors in label semantics, assuming label locations to be either correct or irrelevant (e.g., identical and independently distributed). A few works exist on label registration errors, but these methods often focus on label misalignment on object segment boundaries at the pixel level without guaranteeing vector continuity. To fill the gap, this paper proposes a spatial learning framework based on Expectation-Maximization that iteratively updates deep neural network parameters while inferring true vector label locations. Specifically, inference of true vector locations is based on both the current pixel class predictions and the geometric properties of vectors. Evaluations on real-world high-resolution remote sensing datasets in National Hydrography Dataset (NHD) refinement show that the proposed framework outperforms baseline methods in classification accuracy and refined vector quality.