arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2309.03367cs.CVcs.AI
收藏

面向低资源下游任务的自监督掩码数字高程模型编码

Self-Supervised Masked Digital Elevation Models Encoding for Low-Resource Downstream Tasks

National Center for Supercomputing Applications(国家超级计算应用中心)

Priyam Mazumdar, Aiman Soliman, Volodymyr Kindratenko, Luigi Marini, Kenton McHenry

AI总结:

针对地球表面标注数据匮乏的问题,提出基于ImageNet预训练的Masked Autoencoder与UperNet架构,从数字高程模型中提取建筑物和道路分割,在仅用0.5%-5%数据下取得高IoU,验证了自监督方法在低资源下游任务中的有效性。

摘要:

缺乏高质量的标注数据是训练深度学习模型的主要瓶颈之一。随着任务复杂度的增加,过拟合和不稳定学习的风险也随之上升。目前采用的典型范式是自监督学习,即模型尝试从大量非结构化和无标签的数据中学习,然后将这些知识迁移到目标任务中。其他模态中自监督学习的著名例子包括用于大型语言模型的BERT、用于语音识别的Wav2Vec以及用于视觉的Masked AutoEncoder,它们都利用Transformers来解决掩码预测任务。由于已经收集了数十年的数据,其中很少部分有精确可靠的标注,GeoAI在利用自监督方法方面具有独特优势。我们的目标是从提供地球表面详细地形的数字高程模型(DEM)中提取建筑物和道路分割。提出的架构是在ImageNet上预训练的Masked Autoencoder(但存在ImageNet和DEM之间域差异较大的限制),并使用UperNet Head进行分割解码。我们仅用450和50张训练图像测试了该模型,分别约占原始数据的5%和0.5%。在建筑物分割任务中,该模型用450张图像获得了82.1%的交并比(IoU),仅用50张图像获得了69.1%的IoU。在更具挑战性的道路检测任务中,该模型用450张图像获得了82.7%的IoU,仅用50张图像获得了73.2%的IoU。由于地表景观不断变化的特性,现今任何关于地球表面的人工标注数据集都会很快过时。这促使了能够用于各种下游任务的数据高效学习器的迫切需求。

英文摘要:

The lack of quality labeled data is one of the main bottlenecks for training Deep Learning models. As the task increases in complexity, there is a higher penalty for overfitting and unstable learning. The typical paradigm employed today is Self-Supervised learning, where the model attempts to learn from a large corpus of unstructured and unlabeled data and then transfer that knowledge to the required task. Some notable examples of self-supervision in other modalities are BERT for Large Language Models, Wav2Vec for Speech Recognition, and the Masked AutoEncoder for Vision, which all utilize Transformers to solve a masked prediction task. GeoAI is uniquely poised to take advantage of the self-supervised methodology due to the decades of data collected, little of which is precisely and dependably annotated. Our goal is to extract building and road segmentations from Digital Elevation Models (DEM) that provide a detailed topography of the earths surface. The proposed architecture is the Masked Autoencoder pre-trained on ImageNet (with the limitation that there is a large domain discrepancy between ImageNet and DEM) with an UperNet Head for decoding segmentations. We tested this model with 450 and 50 training images only, utilizing roughly 5% and 0.5% of the original data respectively. On the building segmentation task, this model obtains an 82.1% Intersection over Union (IoU) with 450 Images and 69.1% IoU with only 50 images. On the more challenging road detection task the model obtains an 82.7% IoU with 450 images and 73.2% IoU with only 50 images. Any hand-labeled dataset made today about the earths surface will be immediately obsolete due to the constantly changing nature of the landscape. This motivates the clear necessity for data-efficient learners that can be used for a wide variety of downstream tasks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑