Few-Shot Image Classification

The ImageNet dataset contains 14,197,122 annotated images according to the WordNet hierarchy. Since 2010 the dataset is used in the ImageNet Large Scale Visual Recognition Challenge (ILSVRC), a benchmark in image classification and object detection. The publicly released dataset contains a set of manually annotated training images. A set of test images is also released, with the manual annotations withheld. ILSVRC annotations fall into one of two categories: (1) image-level annotation of a binary label for the presence or absence of an object class in the image, e.g., “there are cars in this image” but “there are no tigers,” and (2) object-level annotation of a tight bounding box and class label around an object instance in the image, e.g., “there is a screwdriver centered at position (20,25) with width of 50 pixels and height of 30 pixels”. The ImageNet project does not own the copyright of the images, therefore only thumbnails and URLs of images are provided. * Total number of non-empty WordNet synsets: 21841 * Total number of images: 14197122 * Number of images with bounding box annotations: 1,034,908 * Number of synsets with SIFT features: 1000 * Number of images with SIFT features: 1.2 million Source: ImageNet Large Scale Visual Recognition Challenge Image Source: https://cs.stanford.edu/people/karpathy/cnnembed/

The CIFAR-100 dataset (Canadian Institute for Advanced Research, 100 classes) is a subset of the Tiny Images dataset and consists of 60000 32x32 color images. The 100 classes in the CIFAR-100 are grouped into 20 superclasses. There are 600 images per class. Each image comes with a "fine" label (the class to which it belongs) and a "coarse" label (the superclass to which it belongs). There are 500 training images and 100 testing images per class. The criteria for deciding whether an image belongs to a class were as follows: * The class name should be high on the list of likely answers to the question “What is in this picture?” * The image should be photo-realistic. Labelers were instructed to reject line drawings. * The image should contain only one prominent instance of the object to which the class refers. * The object may be partially occluded or seen from an unusual viewpoint as long as its identity is still clear to the labeler. Source: https://www.cs.toronto.edu/~kriz/cifar.html Image Source: https://www.cs.toronto.edu/~kriz/cifar.html

CelebFaces Attributes dataset contains 202,599 face images of the size 178×218 from 10,177 celebrities, each annotated with 40 binary labels indicating facial attributes like hair color, gender and age. Source: Show, Attend and Translate: Unpaired Multi-Domain Image-to-Image Translation with Visual Attention Image Source: http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html

The Caltech-UCSD Birds-200-2011 (CUB-200-2011) dataset is the most widely-used dataset for fine-grained visual categorization task. It contains 11,788 images of 200 subcategories belonging to birds, 5,994 for training and 5,794 for testing. Each image has detailed annotations: 1 subcategory label, 15 part locations, 312 binary attributes and 1 bounding box. The textual information comes from Reed et al.. They expand the CUB-200-2011 dataset by collecting fine-grained natural language descriptions. Ten single-sentence descriptions are collected for each image. The natural language descriptions are collected through the Amazon Mechanical Turk (AMT) platform, and are required at least 10 words, without any information of subcategories and actions. Source: Fine-grained Visual-textual Representation Learning Image Source: http://www.vision.caltech.edu/visipedia/CUB-200-2011.html

mini-Imagenet is proposed by Matching Networks for One Shot Learning . In NeurIPS, 2016. This dataset consists of 50000 training images and 10000 testing images, evenly distributed across 100 classes.

The Kinetics dataset is a large-scale, high-quality dataset for human action recognition in videos. The dataset consists of around 500,000 video clips covering 600 human action classes with at least 600 video clips for each action class. Each video clip lasts around 10 seconds and is labeled with a single action class. The videos are collected from YouTube. Source: Self-supervised Visual Feature Learning with Deep Neural Networks: A Survey

Oxford 102 Flower is an image classification dataset consisting of 102 flower categories. The flowers chosen to be flower commonly occurring in the United Kingdom. Each class consists of between 40 and 258 images. The images have large scale, pose and light variations. In addition, there are categories that have large variations within the category and several very similar categories.

The Stanford Cars dataset consists of 196 classes of cars with a total of 16,185 images, taken from the rear. The data is divided into almost a 50-50 train/test split with 8,144 training images and 8,041 testing images. Categories are typically at the level of Make, Model, Year. The images are 360×240. Source: View Independent Vehicle Make, Model and Color Recognition Using Convolutional Neural Network Image Source: https://ai.stanford.edu/~jkrause/cars/car_dataset.html

The iNaturalist 2017 dataset (iNat) contains 675,170 training and validation images from 5,089 natural fine-grained categories. Those categories belong to 13 super-categories including Plantae (Plant), Insecta (Insect), Aves (Bird), Mammalia (Mammal), and so on. The iNat dataset is highly imbalanced with dramatically different number of images per category. For example, the largest super-category “Plantae (Plant)” has 196,613 images from 2,101 categories; whereas the smallest super-category “Protozoa” only has 381 images from 4 categories. Source: Large Scale Fine-Grained Categorization and Domain-Specific Transfer Learning Image Source: https://github.com/visipedia/inat_comp/tree/master/2017

Caltech-256 is an object recognition dataset containing 30,607 real-world images, of different sizes, spanning 257 classes (256 object classes and an additional clutter class). Each class is represented by at least 80 images. The dataset is a superset of the Caltech-101 dataset. Source: Exploiting Non-Linear Redundancy for Neural Model Compression Image Source: ML4A

The tieredImageNet dataset is a larger subset of ILSVRC-12 with 608 classes (779,165 images) grouped into 34 higher-level nodes in the ImageNet human-curated hierarchy. This set of nodes is partitioned into 20, 6, and 8 disjoint sets of training, validation, and testing nodes, and the corresponding classes form the respective meta-sets. As argued in Ren et al. (2018), this split near the root of the ImageNet hierarchy results in a more challenging, yet realistic regime with test classes that are less similar to training classes. Source: tieredImageNet Image Source: https://arxiv.org/pdf/1803.00676.pdf

Animals with Attributes (AwA) was a dataset for benchmarking transfer-learning algorithms, in particular attribute base classification. It consisted of 30475 images of 50 animals classes with six pre-extracted feature representations for each image. The animals classes are aligned with Osherson's classical class/attribute matrix, thereby providing 85 numeric attribute values for each class. Using the shared attributes, it is possible to transfer information between different classes. The Animals with Attributes dataset was suspended. Its images are not available anymore because of copyright restrictions. A drop-in replacement, Animals with Attributes 2, is available instead. Source: Transductive Multi-view Zero-Shot Learning Image Source: https://cvml.ist.ac.at/AwA/

Animals with Attributes 2 (AwA2) is a dataset for benchmarking transfer-learning algorithms, such as attribute base classification and zero-shot learning. AwA2 is a drop-in replacement of original Animals with Attributes (AwA) dataset, with more images released for each category. Specifically, AwA2 consists of in total 37322 images distributed in 50 animal categories. The AwA2 also provides a category-attribute matrix, which contains an 85-dim attribute vector (e.g., color, stripe, furry, size, and habitat) for each category. Source: Learning from Noisy Web Data with Category-level Supervision Image Source: https://arxiv.org/pdf/1604.00326.pdf

ImageNet Long-Tailed is a subset of /dataset/imagenet dataset consisting of 115.8K images from 1000 categories, with maximally 1280 images per class and minimally 5 images per class. The additional classes of images in ImageNet-2010 are used as the open set. Source: Large-Scale Long-Tailed Recognition in an Open World

CIFAR100 few-shots (CIFAR-FS) is randomly sampled from CIFAR-100 (Krizhevsky & Hinton, 2009) by using the same criteria with which miniImageNet has been generated. The average inter-class similarity is sufficiently high to represent a challenge for the current state of the art. Moreover, the limited original resolution of 32×32 makes the task harder and at the same time allows fast prototyping. Source: Bertinetto et al. Image source: Bertinetto et al.

aPY is a coarse-grained dataset composed of 15339 images from 3 broad categories (animals, objects and vehicles), further divided into a total of 32 subcategories (aeroplane, …, zebra). Source: From Classical to Generalized Zero-Shot Learning: a Simple Adaptation Process Image Source: https://www.cs.cmu.edu/~afarhadi/papers/Attributes.pdf

The FC100 dataset (Fewshot-CIFAR100) is a newly split dataset based on CIFAR-100 for few-shot learning. It contains 20 high-level categories which are divided into 12, 4, 4 categories for training, validation and test. There are 60, 20, 20 low-level classes in the corresponding split containing 600 images of size 32 × 32 per class. Smaller image size makes it more challenging for few-shot learning. Source: Prototype Rectification for Few-Shot Learning

The Meta-Dataset benchmark is a large few-shot learning benchmark and consists of multiple datasets of different data distributions. It does not restrict few-shot tasks to have fixed ways and shots, thus representing a more realistic scenario. It consists of 10 datasets from diverse domains: * ILSVRC-2012 (the ImageNet dataset, consisting of natural images with 1000 categories) * Omniglot (hand-written characters, 1623 classes) * Aircraft (dataset of aircraft images, 100 classes) * CUB-200-2011 (dataset of Birds, 200 classes) * Describable Textures (different kinds of texture images with 43 categories) * Quick Draw (black and white sketches of 345 different categories) * Fungi (a large dataset of mushrooms with 1500 categories) * VGG Flower (dataset of flower images with 102 categories), * Traffic Signs (German traffic sign images with 43 classes) * MSCOCO (images collected from Flickr, 80 classes). All datasets except Traffic signs and MSCOCO have a training, validation and test split (proportioned roughly into 70%, 15%, 15%). The datasets Traffic Signs and MSCOCO are reserved for testing only. Source: Optimized Generic Feature Learning for Few-shot Classification across Domains Image Source: Triantafillou et al

Extended GTEA Gaze+ EGTEA Gaze+ is a large-scale dataset for FPV actions and gaze. It subsumes GTEA Gaze+ and comes with HD videos (1280x960), audios, gaze tracking data, frame-level action annotations, and pixel-level hand masks at sampled frames. Specifically, EGTEA Gaze+ contains 28 hours (de-identified) of cooking activities from 86 unique sessions of 32 subjects. These videos come with audios and gaze tracking (30Hz). We have further provided human annotations of actions (human-object interactions) and hand masks. The action annotations include 10325 instances of fine-grained actions, such as "Cut bell pepper" or "Pour condiment (from) condiment container into salad". The hand annotations consist of 15,176 hand masks from 13,847 frames from the videos. Source: http://cbs.ic.gatech.edu/fpv/ Image Source: http://cbs.ic.gatech.edu/fpv/

Places-LT has an imbalanced training set with 62,500 images for 365 classes from Places-2. The class frequencies follow a natural power law distribution with a maximum number of 4,980 images per class and a minimum number of 5 images per class. The validation and testing sets are balanced and contain 20 and 100 images per class respectively. Source: Long-Tailed Recognition Using Class-Balanced Experts

The Stanford Dogs dataset contains 20,580 images of 120 classes of dogs from around the world, which are divided into 12,000 images for training and 8,580 images for testing. Source: Universal-to-Specific Framework for Complex Action Recognition Image Source: https://www.tensorflow.org/datasets/catalog/stanford_dogs

UT Zappos50K is a large shoe dataset consisting of 50,025 catalog images collected from Zappos.com. The images are divided into 4 major categories — shoes, sandals, slippers, and boots — followed by functional types and individual brands. The shoes are centered on a white background and pictured in the same orientation for convenient analysis. Source: UT Zappos50K

When glancing at a magazine, or browsing the Internet, we are continuously being exposed to photographs. Despite of this overflow of visual information, humans are extremely good at remembering thousands of pictures along with some of their visual details. But not all images are equal in memory. Some stitch to our minds, and other are forgotten. In this paper we focus on the problem of predicting how memorable an image will be. We show that memorability is a stable property of an image that is shared across different viewers. We introduce a database for which we have measured the probability that each picture will be remembered after a single view. We analyze image features and labels that contribute to making an image memorable, and we train a predictor based on global image descriptors. We find that predicting image memorability is a task that can be addressed with current computer vision techniques. Whereas making memorable images is a challenging task in visualization and photography, this work is a first attempt to quantify this useful quality of images.

Animal Kingdom is a large and diverse dataset that provides multiple annotated tasks to enable a more thorough understanding of natural animal behaviors. The wild animal footage used in the dataset records different times of the day in an extensive range of environments containing variations in backgrounds, viewpoints, illumination and weather conditions. More specifically, the dataset contains 50 hours of annotated videos to localize relevant animal behavior segments in long videos for the video grounding task, 30K video sequences for the fine-grained multi-label action recognition task, and 33K frames for the pose estimation task, which correspond to a diverse range of animals with 850 species across 6 major animal classes.

Bongard-HOI testifies to which extent your few-shot visual learner can quickly induce the true HOI concept from a handful of images and perform reasoning with it. Further, the learner is also expected to transfer the learned few-shot skills to novel HOI concepts compositionally.

The COCO-MLT is created from MS COCO-2017, containing 1,909 images from 80 classes. The maximum of training number per class is 1,128 and the minimum is 6. We use the test set of COCO2017 with 5,000 for evaluation. The ratio of head, medium, and tail classes is 22:33:25 in COCO-MLT.

We construct the long-tailed version of VOC from its 2012 train-val set. It contains 1,142 images from 20 classes, with a maximum of 775 images per class and a minimum of 4 images per class. The ratio of head, medium, and tail classes after splitting is 6:6:8. We evaluate the performance on VOC2007 test set with 4952 images.

We introduce ArtBench-10, the first class-balanced, high-quality, cleanly annotated, and standardized dataset for benchmarking artwork generation. It comprises 60,000 images of artwork from 10 distinctive artistic styles, with 5,000 training images and 1,000 testing images per style. ArtBench-10 has several advantages over previous artwork datasets. Firstly, it is class-balanced while most previous artwork datasets suffer from the long tail class distributions. Secondly, the images are of high quality with clean annotations. Thirdly, ArtBench-10 is created with standardized data collection, annotation, filtering, and preprocessing procedures. We provide three versions of the dataset with different resolutions (32×32, 256×256, and original image size), formatted in a way that is easy to be incorporated by popular machine learning frameworks.

ORBIT is a real-world few-shot dataset and benchmark grounded in a real-world application of teachable object recognizers for people who are blind/low vision. The dataset contains 3,822 videos of 486 objects recorded by people who are blind/low-vision on their mobile phones, and the benchmark reflects a realistic, highly challenging recognition problem, providing a rich playground to drive research in robustness to few-shot, high-variation conditions.

NIH-CXR-LT. NIH ChestXRay14 contains over 100,000 chest X-rays labeled with 14 pathologies, plus a “No Findings” class. We construct a single-label, long-tailed version of the NIH ChestXRay14 dataset by introducing five new disease findings described above. The resulting NIH-CXR-LT dataset has 20 classes, including 7 head classes, 10 medium classes, and 3 tail classes. NIH-CXR-LT contains 88,637 images labeled with one of 19 thorax diseases, with 68,058 training and 20,279 test images. The validation and balanced test sets contain 15 and 30 images per class, respectively.

The Few-Shot Object Learning (FewSOL) dataset can be used for object recognition with a few images per object. It contains 336 real-world objects with 9 RGB-D images per object from different views. Object segmentation masks, object poses and object attributes are provided. In addition, synthetic images generated using 330 3D object models are used to augment the dataset. FewSOL dataset can be used to study a set of few-shot object recognition problems such as classification, detection and segmentation, shape reconstruction, pose estimation, keypoint correspondences and attribute recognition. Motivation: If robots can recognize objects from a few exemplar images, it is possible to scale up the number of objects a robot can recognize because collecting a few images per object is a much easier process compared to building a 3D model of an object. In addition, models trained in the meta-learning setting can generalize to new objects without re-training.

MIMIC-CXR-LT. We construct a single-label, long-tailed version of MIMIC-CXR in a similar manner. MIMIC-CXR is a multi-label classification dataset with over 200,000 chest X-rays labeled with 13 pathologies and a “No Findings” class. The resulting MIMIC-CXR-LT dataset contains 19 classes, of which 10 are head classes, 6 are medium classes, and 3 are tail classes. MIMIC-CXR-LT contains 111,792 images labeled with one of 18 diseases, with 87,493 training images and 23,550 test set images. The validation and balanced test sets contain 15 and 30 images per class, respectively.

Geoclidean-Constraints dataset consists of 20 concepts and 40 tasks, created from permutations of line and circle construction rules with various constraints describing the relationship between objects. This dataset focuses on explicit constraints between geometric objects. We denote the objects as the following—lines as L, circles as C, and triangles (constructed from three lines) as T. Source: Geoclidean: Few-Shot Generalization in Euclidean Geometry Image Source: https://arxiv.org/pdf/2211.16663v1.pdf

mini-ImageNet was proposed by Matching networks for one-shot learning for few-shot learning evaluation, in an attempt to have a dataset like ImageNet while requiring fewer resources. Similar to the statistics for CIFAR-100-LT with an imbalance factor of 100, we construct a long-tailed variant of mini-ImageNet that features all the 100 classes and an imbalanced training set with $N_1 = 500$ and $N_K = 5$ images. For evaluation, both the validation and test sets are balanced and contain 10K images, 100 samples for each of the 100 categories.

The development of the remote sensing fine-grained ship classification field necessitates large-scale realistic fine-grained ship datasets. The FGSC-23 and FGSCR-42 datasets have been pivotal for practical applications, yet they exhibit limitations when assessing advanced classification methods. Specifically, FGSC-23 categorizes data into relatively broad classes, such as lumping all auxiliary ships into one category, which does not align with the granularity required in real-world scenarios. Additionally, the dataset's limited number of categories fails to satisfy the diverse needs of practical applications. FGSCR-42, while providing finer classifications within some broader categories, still lacks comprehensive subcategory coverage. Moreover, the dataset suffers from an imbalance in sample distribution, with six categories containing fewer than ten samples each, which can hinder effective model training. To overcome these limitations, we introduce the FGSCM-52 dataset, an enhanced version of FGSCR-42, designed to offer a more extensive collection of fine-grained remote sensing ship images. FGSCM-52 includes an additional ten categories and increases the number of samples across existing categories, thereby making it serve as a better testbed for fine-grained ship classification.The dataset features image sizes ranging from 50$\times$50 to 1,600$\times$1,700 pixels. Furthermore, FGSCM-52 is equipped with well-defined hierarchical labels, which facilitate a more nuanced understanding of ship classification.