当前位置: X-MOL 学术IEEE Multimed. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Learning Quintuplet Loss for Large-Scale Visual Geolocalization
IEEE Multimedia ( IF 2.3 ) Pub Date : 2020-05-25 , DOI: 10.1109/mmul.2020.2996941
Qiang Zhai 1 , Rui Huang 1 , Hong Cheng 1 , Huiqin Zhan 1 , Jun Li 2 , Zicheng Liu 3
Affiliation  

With the maturity of artificial intelligence technology, large-scale visual geolocalization (LSVGL) is increasingly important in urban computing, where the task is to accurately and efficiently recognize the geolocation of a given query image. The main challenge of LSVGL faced by many experiments due to the appearance of real-word places may differ in various ways while perspective deviation almost inevitably exists between training images and query images because of the arbitrary perspective. To cope with this situation, in this article, we in-depth analyze the limitation of triplet loss, which is the most commonly used metric learning loss in state-of-the-art LSVGL framework and propose a new quintuplet loss by embedding all the potential positive samples to the primitive triplet loss. Extensive experiments are conducted to verify the effectiveness of the proposed approach and the results demonstrate that our new loss can enhance various LSVGL methods.

中文翻译:


学习大规模视觉地理定位的五元组损失



随着人工智能技术的成熟,大规模视觉地理定位(LSVGL)在城市计算中变得越来越重要,其任务是准确有效地识别给定查询图像的地理位置。许多实验面临的主要挑战是,由于真实世界地点的出现可能会以各种方式存在差异,而由于任意视角,训练图像和查询图像之间几乎不可避免地存在视角偏差。为了应对这种情况,在本文中,我们深入分析了三元组损失(三元组损失)的局限性,三元组损失是最先进的LSVGL框架中最常用的度量学习损失,并提出了一种新的五元组损失,通过嵌入所有原始三重态损失的潜在正样本。进行了大量的实验来验证所提出方法的有效性,结果表明我们的新损失可以增强各种 LSVGL 方法。
更新日期:2020-05-25
down
wechat
bug