New York Smells: A Large Multimodal Dataset for Olfaction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ozguroglu, Ege, Liang, Junbang, Liu, Ruoshi, Chiquier, Mia, DeTienne, Michael, Qian, Wesley Wei, Horowitz, Alexandra, Owens, Andrew, Vondrick, Carl
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918217991061504
author Ozguroglu, Ege
Liang, Junbang
Liu, Ruoshi
Chiquier, Mia
DeTienne, Michael
Qian, Wesley Wei
Horowitz, Alexandra
Owens, Andrew
Vondrick, Carl
author_facet Ozguroglu, Ege
Liang, Junbang
Liu, Ruoshi
Chiquier, Mia
DeTienne, Michael
Qian, Wesley Wei
Horowitz, Alexandra
Owens, Andrew
Vondrick, Carl
contents While olfaction is central to how animals perceive the world, this rich chemical sensory modality remains largely inaccessible to machines. One key bottleneck is the lack of diverse, multimodal olfactory training data collected in natural settings. We present New York Smells, a large dataset of paired image and olfactory signals captured ``in the wild.'' Our dataset contains 7,000 smell-image pairs from 3,500 distinct objects across indoor and outdoor environments, with approximately 70$\times$ more objects than existing olfactory datasets. Our benchmark has three tasks: cross-modal smell-to-image retrieval, recognizing scenes, objects, and materials from smell alone, and fine-grained discrimination between grass species. Through experiments on our dataset, we find that visual data enables cross-modal olfactory representation learning, and that our learned olfactory representations outperform widely-used hand-crafted features.
format Preprint
id arxiv_https___arxiv_org_abs_2511_20544
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle New York Smells: A Large Multimodal Dataset for Olfaction
Ozguroglu, Ege
Liang, Junbang
Liu, Ruoshi
Chiquier, Mia
DeTienne, Michael
Qian, Wesley Wei
Horowitz, Alexandra
Owens, Andrew
Vondrick, Carl
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
While olfaction is central to how animals perceive the world, this rich chemical sensory modality remains largely inaccessible to machines. One key bottleneck is the lack of diverse, multimodal olfactory training data collected in natural settings. We present New York Smells, a large dataset of paired image and olfactory signals captured ``in the wild.'' Our dataset contains 7,000 smell-image pairs from 3,500 distinct objects across indoor and outdoor environments, with approximately 70$\times$ more objects than existing olfactory datasets. Our benchmark has three tasks: cross-modal smell-to-image retrieval, recognizing scenes, objects, and materials from smell alone, and fine-grained discrimination between grass species. Through experiments on our dataset, we find that visual data enables cross-modal olfactory representation learning, and that our learned olfactory representations outperform widely-used hand-crafted features.
title New York Smells: A Large Multimodal Dataset for Olfaction
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2511.20544