Google Landmarks Dataset v2 – A Large-Scale Benchmark for Instance-Level Recognition and Retrieval

By Tobias Weyand, Andre F. de Araújo, Bingyi Cao, Jack Sim. 2020

477 citations recorded by the source corpus.

While image retrieval and instance recognition techniques are progressing rapidly, there is a need for challenging datasets to accurately measure their performance -- while posing novel challenges that are relevant for practical applications. We introduce the Google Landmarks Dataset v2 (GLDv2), a new benchmark for large-scale, fine-grained instance recognition and image retrieval in the domain of human-made and natural landmarks. GLDv2 is the largest such dataset to date by a large margin, including over 5M images and 200k distinct instance labels. Its test set consists of 118k images with ground truth annotations for both the retrieval and recognition tasks. The ground truth construction involved over 800 hours of human annotator work. Our new dataset has several challenging properties inspired by real-world applications that previous datasets did not consider: An extremely long-tailed class distribution, a large fraction of out-of-domain test photos and large intra-class variability. The dataset is sourced from Wikimedia Commons, the world's largest crowdsourced collection of landmark photos. We provide baseline results for both recognition and retrieval tasks based on state-of-the-art methods as well as competitive results from a public challenge. We further demonstrate the suitability of the dataset for transfer learning by showing that image embeddings trained on it achieve competitive retrieval performance on independent datasets. The dataset images, ground-truth and metric scoring code are available at https://github.com/cvdfoundation/google-landmark

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Evaluation & Benchmarks

Citations in the corpus

Are we done with ImageNet? (2020, background): “Several studies have revisited common computer vision benchmarks [16, 17, 18, 19] and, regarding ImageNet [1] speciﬁcally, identiﬁed various sources of bias and noise [20, 21, 22].”
Emerging Properties in Self-Supervised Vision Transformers (2021, neutral): “2M clean set from Google Landmarks v2 (GLDv2) [59], a dataset of landmarks designed for retrieval purposes.”
Federated Visual Classification with Real-World Data Distribution (2020, methodology): “Google Landmarks Dataset V2 (GLD-v2) [32] is a large scale image dataset for landmark recognition and retrieval, consisting of 5 million images with each attributed to one of over 280,000 authors.”
Metric Compatible Training for Online Backfilling in Large-Scale Retrieval (2023, unknown): “Open-class setting We apply the proposed approach to an open-class setting, where the old and new data do not share the same classes, on CIFAR-100 and Google Landmarks V2 (GLDv2) [25] with the ResNet-18 backbone network.”
3rd Continual Learning Workshop Challenge on Egocentric Category and Instance Level Object Understanding (2022, background): “Apart from objects, datasets for instance-level recognition exist for places [49], art [50], and clothing [51].”
Boundary-aware Backward-Compatible Representation via Adversarial Learning in Image Retrieval (2023, methodology): “Researchers pay more attention to the improvement of retrieval performance [3, 13, 25, 27, 28] and there exist several comprehensive benchmarks [12, 22].”
COFAR: Commonsense and Factual Reasoning in Image Search (2022, background): “2016) and the Google landmarks (Weyand et al., 2020) respectively.”
Product1M: Towards Weakly Supervised Instance-Level Product Retrieval via Cross-Modal Pretraining (2021, methodology): “We adopt Precision (Prec@N ), mean Average Precision (mAP@N ) and mean Average Recall (mAR@N ) as our evaluation metrics, among which Prec@N and mAP@N are widely used in the retrieval literature [49, 3].”
Open X-Embodiment: Robotic Learning Datasets and RT-X Models : Open X-Embodiment Collaboration0 (2023, unknown): “Even the largest data collection efforts still end up with datasets that are a fraction of the size and diversity of benchmark datasets in vision (5-18M) [4, 5] and NLP (1.5B-4.”
MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models (2023, background): “Similar with the above coarse-grained recognition, the images of these subtasks are from publicly available datasets [25, 37, 38, 55, 46] and all of the instructions are manually designed.”
Deep Class-Incremental Learning: A Survey (2023, methodology): “There are other works [75], [77], [82], [92], [104], [116], [139] utilizing MNIST [191], Core50 [192], Google Landmarkv2 [193], CUB200-2011 [194], TinyImageNet [195], and miniImageNet [190] for model evaluation, while the aforementioned datasets are the most widely adopted in the current CIL communi”
Learning and aggregating deep local descriptors for instance-level recognition (2020, methodology): “The Google Landmarks Dataset – version 2 (GLD2) [48] is used.”
Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Models (2022, methodology): “GPR1200 is a general-purpose content retrieval dataset with 1200 classes sampled from other datasets such as Google Landmarks V2 [65], Stanford Online Products [42], IMDB-WIKI [55] and others.”
Unifying Deep Local and Global Features for Image Search (2020, methodology): “Note that recent work has reported even higher performance on the retrieval task, by learning on GLDv2’s training set and using query expansion techniques [64] / ensembling [62].”