# Enhancing Remote Sensing Image Retrieval with Triplet Deep Metric Learning Network

arXiv: https://arxiv.org/abs/1902.05818 · HTML: https://mlsanity.com/paper/1902.05818/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 15 Feb 2019.

## Built on by (most cited first, 12 of 12)

- [Composed Image Retrieval for Remote Sensing](https://mlsanity.com/paper/2405.15587/) (2024)
- [Deep Unsupervised Contrastive Hashing for Large-Scale Cross-Modal Text-Image Retrieval in Remote Sensing](https://mlsanity.com/paper/2201.08125/) (2022)
- [Informative and Representative Triplet Selection for Multilabel Remote Sensing Image Retrieval](https://mlsanity.com/paper/2105.03647/) (2021)
- [Unsupervised Contrastive Hashing for Cross-Modal Retrieval in Remote Sensing](https://mlsanity.com/paper/2204.08707/) (2022)
- [Airplane Detection Based on Mask Region Convolution Neural Network](https://mlsanity.com/paper/2108.12817/) (2021)
- [Object-Level Targeted Selection via Deep Template Matching](https://mlsanity.com/paper/2207.01778/) (2022)
- [A Triplet-loss Dilated Residual Network for High-Resolution Representation Learning in Image Retrieval](https://mlsanity.com/paper/2303.08398/) (2023)
- [Creating Ensembles of Classifiers through UMDA for Aerial Scene Classification](https://mlsanity.com/paper/2303.11389/) (2023)
- [Annotation Cost-Efficient Active Learning for Deep Metric Learning Driven Remote Sensing Image Retrieval](https://mlsanity.com/paper/2406.10107/) (2024)
- [GLRT-Based Metric Learning for Remote Sensing Object Retrieval](https://mlsanity.com/paper/2410.05773/) (2024)
- [MARC: Multi-Label Adaptive Retrieval Contrastive Loss for Remote Sensing Images](https://mlsanity.com/paper/2512.16294/) (2025)
- [Benchmarking Composed Image Retrieval for Applied Earth Observation](https://mlsanity.com/paper/2605.24442/) (2026)

## Abstract (arXiv)

With the rapid growing of remotely sensed imagery data, there is a high demand for effective and efficient image retrieval tools to manage and exploit such data. In this letter, we present a novel content-based remote sensing image retrieval method based on Triplet deep metric learning convolutional neural network (CNN). By constructing a Triplet network with metric learning objective function, we extract the representative features of the images in a semantic space in which images from the same class are close to each other while those from different classes are far apart. In such a semantic space, simple metric measures such as Euclidean distance can be used directly to compare the similarity of images and effectively retrieve images of the same class. We also investigate a supervised and an unsupervised learning methods for reducing the dimensionality of the learned semantic features. We present comprehensive experimental results on two publicly available remote sensing image retrieval datasets and show that our method significantly outperforms state-of-the-art.