Image Classification

The SNIPS Natural Language Understanding benchmark is a dataset of over 16,000 crowdsourced queries distributed among 7 user intents of various complexity: * SearchCreativeWork (e.g. Find me the I, Robot television show), * GetWeather (e.g. Is it windy in Boston, MA right now?), * BookRestaurant (e.g. I want to book a highly rated restaurant in Paris tomorrow night), * PlayMusic (e.g. Play the last track from Beyoncé off Spotify), * AddToPlaylist (e.g. Add Diamonds to my roadtrip playlist), * RateBook (e.g. Give 6 stars to Of Mice and Men), * SearchScreeningEvent (e.g. Check the showtimes for Wonder Woman in Paris). The training set contains of 13,084 utterances, the validation set and the test set contain 700 utterances each, with 100 queries per intent. Source: https://paperswithcode.com/paper/snips-voice-platform-an-embedded-spoken/

YFCC100M is a that dataset contains a total of 100 million media objects, of which approximately 99.2 million are photos and 0.8 million are videos, all of which carry a Creative Commons license. Each media object in the dataset is represented by several pieces of metadata, e.g. Flickr identifier, owner name, camera, title, tags, geo, media source. The collection provides a comprehensive snapshot of how photos and videos were taken, described, and shared over the years, from the inception of Flickr in 2004 until early 2014. Source: YFCC100M: The New Data in Multimedia Research Image Source: Thomee et al

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

Stanford Online Products (SOP) dataset has 22,634 classes with 120,053 product images. The first 11,318 classes (59,551 images) are split for training and the other 11,316 (60,502 images) classes are used for testing Source: Deep Metric Learning with Alternating Projections onto Feasible Sets Image Source: https://cvgl.stanford.edu/projects/lifted_struct/

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

Consists of more than 210k videos for 310 audio classes. Source: VGGSound: A Large-scale Audio-Visual Dataset

AI2 Diagrams (AI2D) is a dataset of over 5000 grade school science diagrams with over 150000 rich annotations, their ground truth syntactic parses, and more than 15000 corresponding multiple choice questions. Source: A Diagram Is Worth A Dozen Images Image Source: Kembhavi et al

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.

The Visual Task Adaptation Benchmark (VTAB) is a benchmark designed to evaluate general visual representations². It consists of a diverse and challenging suite of tasks². The benchmark defines a good general visual representation as one that yields good performance on unseen tasks, when trained on limited task-specific data². The VTAB benchmark contains the following 19 tasks that are derived from public datasets¹: - Caltech101 - CIFAR-100 - CLEVR distance prediction - CLEVR counting - Diabetic Rethinopathy - Dmlab Frames - dSprites orientation prediction - dSprites location prediction - Describable Textures Dataset (DTD) - EuroSAT - KITTI distance prediction - 102 Category Flower Dataset - Oxford IIIT Pet dataset - PatchCamelyon - Resisc45 - Small NORB azimuth prediction - Small NORB elevation prediction - SUN397 - SVHN The given model is independently fine-tuned for solving each of the above tasks¹. Average accuracy across all tasks is used to measure the model's performance¹. Detailed description of all tasks, evaluation protocol, and other details can be found in the VTAB paper¹. (1) Visual Task Adaptation Benchmark. https://google-research.github.io/task_adaptation/. (2) GitHub - google-research/task_adaptation. https://github.com/google-research/task_adaptation. (3) GitHub - KMnP/vpt: ️ Visual Prompt Tuning [ECCV 2022] https://arxiv .... https://github.com/KMnP/vpt.

CINIC-10 is a dataset for image classification. It has a total of 270,000 images, 4.5 times that of CIFAR-10. It is constructed from two different sources: ImageNet and CIFAR-10. Specifically, it was compiled as a bridge between CIFAR-10 and ImageNet. It is split into three equal subsets - train, validation, and test - each of which contain 90,000 images. Source: Group Knowledge Transfer:Collaborative Training of Large CNNs on the Edge Image Source: https://arxiv.org/abs/1810.03505

RESISC45 dataset is a dataset for Remote Sensing Image Scene Classification (RESISC). It contains 31,500 RGB images of size 256×256 divided into 45 scene classes, each class containing 700 images. Among its notable features, RESISC45 contains varying spatial resolution ranging from 20cm to more than 30m/px.

The Extended Yale B database contains 2414 frontal-face images with size 192×168 over 38 subjects and about 64 images per subject. The images were captured under different lighting conditions and various facial expressions. Source: Learning Locality-Constrained Collaborative Representation for Robust Face Recognition Image Source: http://vision.ucsd.edu/~leekc/ExtYaleDatabase/ExtYaleB.html

The WebVision dataset is designed to facilitate the research on learning visual representation from noisy web data. It is a large scale web images dataset that contains more than 2.4 million of images crawled from the Flickr website and Google Images search. The same 1,000 concepts as the ILSVRC 2012 dataset are used for querying images, such that a bunch of existing approaches can be directly investigated and compared to the models trained from the ILSVRC 2012 dataset, and also makes it possible to study the dataset bias issue in the large scale scenario. The textual information accompanied with those images (e.g., caption, user tags, or description) are also provided as additional meta information. A validation set contains 50,000 images (50 images per category) is provided to facilitate the algorithmic development.

LabelMe database is a large collection of images with ground truth labels for object detection and recognition. The annotations come from two different sources, including the LabelMe online annotation tool. Source: LabelMe: A Database and Web-Based Tool for Image Annotation Image Source: Russell et al

ObjectNet is a test set of images collected directly using crowd-sourcing. ObjectNet is unique as the objects are captured at unusual poses in cluttered, natural scenes, which can severely degrade recognition performance. There are 50,000 images in the test set which controls for rotation, background and viewpoint. There are 313 object classes with 113 overlapping ImageNet. Source: On Robustness and Transferability of Convolutional Neural Networks Image Source: https://objectnet.dev/

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

Functional Map of the World (fMoW) is a dataset that aims to inspire the development of machine learning models capable of predicting the functional purpose of buildings and land use from temporal sequences of satellite images and a rich set of metadata features. Source: Functional Map of the World Image Source: Christie et al

NABirds V1 is a collection of 48,000 annotated photographs of the 400 species of birds that are commonly observed in North America. More than 100 photographs are available for each species, including separate annotations for males, females and juveniles that comprise 700 visual categories. This dataset is to be used for fine-grained visual categorization experiments. Source: https://dl.allaboutbirds.org/nabirds Image Source: https://openaccess.thecvf.com/content_cvpr_2015/papers/Horn_Building_a_Bird_2015_CVPR_paper.pdf

The UCF-Crime dataset is a large-scale dataset of 128 hours of videos. It consists of 1900 long and untrimmed real-world surveillance videos, with 13 realistic anomalies including Abuse, Arrest, Arson, Assault, Road Accident, Burglary, Explosion, Fighting, Robbery, Shooting, Stealing, Shoplifting, and Vandalism. These anomalies are selected because they have a significant impact on public safety. This dataset can be used for two tasks. First, general anomaly detection considering all anomalies in one group and all normal activities in another group. Second, for recognizing each of 13 anomalous activities. Source: Video Anomaly Dection Dataset

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

Oxford5K is the Oxford Buildings Dataset, which contains 5062 images collected from Flickr. It offers a set of 55 queries for 11 landmark buildings, five for each landmark. Source: Unsupervised Adversarial Attacks on Deep Feature-based Retrieval with GAN 1 Corresponding Author Image Source: https://www.robots.ox.ac.uk/~vgg/data/oxbuildings/

The Free Music Archive (FMA) is a large-scale dataset for evaluating several tasks in Music Information Retrieval. It consists of 343 days of audio from 106,574 tracks from 16,341 artists and 14,854 albums, arranged in a hierarchical taxonomy of 161 genres. It provides full-length and high-quality audio, pre-computed features, together with track- and user-level metadata, tags, and free-form text such as biographies. There are four subsets defined by the authors: * Full: the complete dataset, * Large: the full dataset with audio limited to 30 seconds clips extracted from the middle of the tracks (or entire track if shorter than 30 seconds), * Medium: a selection of 25,000 30s clips having a single root genre, * Small: a balanced subset containing 8,000 30s clips with 1,000 clips per one of 8 root genres. The official split into training, validation and test sets (80/10/10) uses stratified sampling to preserve the percentage of tracks per genre. Songs of the same artists are part of one set only. Source: FMA: A Dataset For Music Analysis Audio Source: https://github.com/mdeff/fma

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

PASCAL VOC 2007 is a dataset for image recognition. The twenty object classes that have been selected are: Person: person Animal: bird, cat, cow, dog, horse, sheep Vehicle: aeroplane, bicycle, boat, bus, car, motorbike, train Indoor: bottle, chair, dining table, potted plant, sofa, tv/monitor The dataset can be used for image classification and object detection tasks. Image Source: Object Detection and Recognition in Images

JFT-300M is an internal Google dataset used for training image classification models. Images are labeled using an algorithm that uses complex mixture of raw web signals, connections between web-pages and user feedback. This results in over one billion labels for the 300M images (a single image can have multiple labels). Of the billion image labels, approximately 375M are selected via an algorithm that aims to maximize label precision of selected images.

The KVASIR Dataset was released as part of the medical multimedia challenge presented by MediaEval. It is based on images obtained from the GI tract via an endoscopy procedure. The dataset is composed of images that are annotated and verified by medical doctors, and captures 8 different classes. The classes are based on three anatomical landmarks (z-line, pylorus, cecum), three pathological findings (esophagitis, polyps, ulcerative colitis) and two other classes (dyed and lifted polyps, dyed resection margins) related to the polyp removal process. Overall, the dataset contains 8,000 endoscopic images, with 1,000 image examples per class. Source: Two-Stream Deep Feature Modelling for Automated Video Endoscopy Data Analysis Image Source: https://datasets.simula.no/kvasir/

Imagenet32 is a huge dataset made up of small images called the down-sampled version of Imagenet. Imagenet32 is composed of 1,281,167 training data and 50,000 test data with 1,000 labels. Source: Self-supervised Knowledge Distillation Using Singular Value Decomposition Image Source: https://arxiv.org/pdf/1707.08819v3.pdf

The smallNORB dataset is a datset for 3D object recognition from shape. It contains images of 50 toys belonging to 5 generic categories: four-legged animals, human figures, airplanes, trucks, and cars. The objects were imaged by two cameras under 6 lighting conditions, 9 elevations (30 to 70 degrees every 5 degrees), and 18 azimuths (0 to 340 every 20 degrees). The training set is composed of 5 instances of each category (instances 4, 6, 7, 8 and 9), and the test set of the remaining 5 instances (instances 0, 1, 2, 3, and 5). Source: https://cs.nyu.edu/~ylclab/data/norb-v1.0-small/ Image Source: https://www.kaggle.com/nepuerto/the-small-norb-dataset-v10

The Neuromorphic-Caltech101 (N-Caltech101) dataset is a spiking version of the original frame-based Caltech101 dataset. The original dataset contained both a "Faces" and "Faces Easy" class, with each consisting of different versions of the same images. The "Faces" class has been removed from N-Caltech101 to avoid confusion, leaving 100 object classes plus a background class. The N-Caltech101 dataset was captured by mounting the ATIS sensor on a motorized pan-tilt unit and having the sensor move while it views Caltech101 examples on an LCD monitor as shown in the video below. A full description of the dataset and how it was created can be found in the paper below. Please cite this paper if you make use of the dataset.

PatchCamelyon is an image classification dataset. It consists of 327.680 color images (96 x 96px) extracted from histopathologic scans of lymph node sections. Each image is annotated with a binary label indicating presence of metastatic tissue. PCam provides a new benchmark for machine learning models: bigger than CIFAR10, smaller than ImageNet, trainable on a single GPU.

The RVL-CDIP dataset consists of scanned document images belonging to 16 classes such as letter, form, email, resume, memo, etc. The dataset has 320,000 training, 40,000 validation and 40,000 test images. The images are characterized by low quality, noise, and low resolution, typically 100 dpi. Source: Towards a Multi-modal, Multi-task Learning based Pre-training Framework for Document Representation Learning Image Source: https://www.cs.cmu.edu/~aharley/rvl-cdip/

The Stylized-ImageNet dataset is created by removing local texture cues in ImageNet while retaining global shape information on natural images via AdaIN style transfer. This nudges CNNs towards learning more about shapes and less about local textures. Source: Adversarial Examples Improve Image Recognition Image Source: https://github.com/rgeirhos/Stylized-ImageNet

Comprises 11 hand gesture categories from 29 subjects under 3 illumination conditions. Source: A Low Power, Fully Event-Based Gesture Recognition System

The image dataset TinyImages contains 80 million images of size 32×32 collected from the Internet, crawling the words in WordNet. The authors have decided to withdraw it because it contains offensive content, and have asked the community to stop using it. Source: Implementing Randomized Matrix Algorithms in Parallel and Distributed Environments

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/

This work presents two new benchmark datasets (CIFAR-10N, CIFAR-100N), equipping the training dataset of CIFAR-10 and CIFAR-100 with human-annotated real-world noisy labels that we collect from Amazon Mechanical Turk.