The iNaturalist Species Classification and Detection Dataset

By Grant Van Horn, Oisin Mac Aodha, Yang Song, Yin Cui, Chen Sun, Alexander Shepard, Hartwig Adam, Pietro Perona, Serge J. Belongie. 2017

2,111 citations recorded by the source corpus.

Existing image classification datasets used in computer vision tend to have a uniform distribution of images across object categories. In contrast, the natural world is heavily imbalanced, as some species are more abundant and easier to photograph than others. To encourage further progress in challenging real world conditions we present the iNaturalist species classification and detection dataset, consisting of 859,000 images from over 5,000 different species of plants and animals. It features visually similar species, captured in a wide variety of situations, from all over the world. Images were collected with different camera types, have varying image quality, feature a large class imbalance, and have been verified by multiple citizen scientists. We discuss the collection of the dataset and present extensive baseline experiments using state-of-the-art computer vision classification and detection models. Results show that current non-ensemble based methods achieve only 67% top one classification accuracy, illustrating the difficulty of the dataset. Specifically, we observe poor results for classes with small numbers of training examples suggesting more attention is needed in low-shot learning.

Subfields: Computer Vision and Pattern Recognition, Ecological Modeling, Molecular Biology, Plant Science, 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 Computer Vision and Pattern Recognition

In Ecological Modeling

In Molecular Biology

In Plant Science

In Evaluation & Benchmarks

Citations in the corpus

VirTex: Learning Visual Representations from Textual Annotations (2020, methodology): “We evaluate features learned using VirTex on four downstream tasks with fine-tuning: (a) Instance Segmentation on COCO [30]; (b) Instance Segmentation on LVIS [31]; and (c) Object Detection on PASCAL VOC [99]; (d) Finegrained Classification on iNaturalist 2018 [107].”
Self-supervised Pretraining of Visual Features in the Wild (2021, methodology): “We consider the following benchmarks: iNaturalist [50], OpenImages [31], Places205 [56] and Pascal VOC [16].”
Self-Supervised Learning of Pretext-Invariant Representations (2019, positive): “We follow the evaluation setup from [21] and measure the performance of such classifiers on four image-classification datasets: ImageNet [66], VOC07 [16], Places205 [88], and iNaturalist2018 [74].”
LVIS: A Dataset for Large Vocabulary Instance Segmentation (2019, background): “iNaturalist [26] contains nearly 900k images annotated with bounding boxes for an astonishing 5000 plant and animal species.”
NDD20: A large-scale few-shot dolphin dataset for coarse and fine-grained categorisation (2020, background): “Some large scale datasets showing animals in natural environments do exist (15), although these often only provide labels at a species level, which is not ﬁne-grained enough for population estimation which requires the identiﬁcation of individuals.”
Unsupervised Learning of Visual Features by Contrasting Cluster Assignments (2020, neutral): “First, we report the linear classification performance on the Places205 [58], VOC07 [16], and iNaturalist2018 [46] datasets.”
Differential Privacy Has Disparate Impact on Model Accuracy (2019, methodology): “We use a 60,000-image subset of iNaturalist [19], an 800,000-image dataset of hierarchically labeled plants and animals in natural environments.”
Federated Visual Classification with Real-World Data Distribution (2020, methodology): “Natural Species Classification We create a dataset and classification problem based on the iNaturalist 2017 Challenge [30], where images are contributed by a”
Neural Prototype Trees for Interpretable Fine-grained Image Recognition (2020, methodology): “For CUB, ResNet50 is pretrained on iNaturalist2017 [49], containing plants and animals and therefore a suitable source domain [32], using the backbone of [59].”
Meta-Learning with Less Forgetting on Large-Scale Non-Stationary Task Distributions (2022, neutral): “To investigate the sensitivity of baselines and our method to dataset order, we also performed comparisons on two other dataset sequences, including (i) Butterfly, CUB, CIFARFS, Plantae, MiniIm-agenet, Aircraft and (ii) CUB, CIFARFS, Plantae, MiniImagenet, Butterfly, Aircraft.”
CLiMB: A Continual Learning Benchmark for Vision-and-Language Tasks (2022, background): “Vision-Only Tasks CLiMB includes ImageNet-1000 [Russakovsky et al., 2015], iNatural-ist2019 [Van Horn et al., 2018], Places365 [Mahajan et al., 2018], and MS-COCO object detection (formulated as a multi-label classification task).”
Continual Pre-Training Mitigates Forgetting in Language and Vision (2022, methodology): “The CV environment uses iNaturalist (Horn et al., 2018) for Continual Pre-Training and CORe50 (Lomonaco & Maltoni, 2017) as FC dataset for catastrophic forgetting.”
The Challenges of Continuous Self-Supervised Learning (2022, methodology): “We evaluate these representations by training a linear classifier on the ImageNet [ 16 ] and iNatural-ist [ 76 ] datasets.”
Self-Supervised Training Enhances Online Continual Learning (2021, methodology): “SwAV is the first self-supervised method to surpass ImageNet supervised features by performing a linear evaluation on datasets like Places-205 [74], VOC-07 [22], and iNaturalist-2018 [66].”