6 Mins Read

Important Computer Vision Datasets

computer vision datasets

Important Computer Vision Datasets

One of the most rapidly developing artificial intelligence domains is computer vision. In computer vision, datasets are very important. As a result of the advances made in this field, a large amount of data is now available for training computer vision models. This article reviews some of the most famous CV datasets.

MNIST dataset

MNIST is an introductory computer vision dataset containing 70,000 images of handwritten digits. Data preparation was carried out by Professor Yann Lechon. The data is divided into two sets, a training set of 60,000, and a test set of 10,000 images. All figures are placed in the center of the image.

Computer Vision Datasets
Figure1. The MNIST dataset of handwritten digits. Source

MNIST fashion dataset

This dataset contains grayscale, 28×28 pixels images similar to MNIST images. Various fashion topics are covered, including t-shirts, pants, coats, sandals, blouses, sports shoes, bags, and boots. The data was provided by the Zalando (fashion and clothing store) research team.

computer vision datasets
Figure2. The MNIST fashion dataset. Source

CIFAR-10 and CIFAR-100 datasets

The Canadian Institute for Advanced Research has released both CIFAR-10 and CIFAR-100 datasets. CIFAR-10 consists of 60,000 images in 10 categories including airplanes, cars, birds, cats, deer, dogs, frogs, horses, ships, and trucks. CIFAR-100 is a similar dataset with a total of 60,000 images in 100 categories. Since both datasets are available in 32×32 pixels and have 50,000 training images and 10,000 test images with equally divided data ratios, even beginners can easily use them.

computer vision datasets
Figure3. The CIFAR-10 dataset. Source

 

ImageNet dataset

This dataset was created for a computer vision competition called the “ImageNet Large Scale Visual Recognition Challenge” which challenges teams in five sections. ImageNet is based on the WordNet lexical database and contains over 1.4 million images in over 220,000 categories which makes it the largest collection of publicly available images.

computer vision datasets
Figure4. The Imagenet dataset. Source

PASCAL VOC dataset

This dataset has been made available to the public by Pascal Research Institute. There are four different types of images in this dataset: household images, vehicle images, animal images, and human images divided into 20 object classes. While the PASCAL VOC dataset has fewer categories and numbers than ImageNet, it can be used for a broader range of image segmentation and object detection applications.

computer vision datasets
Figure5. The PASCAL VOC dataset. Source

IMDB-Wiki dataset

This dataset contains 520,000 images of faces taken from IMDB and Wikipedia. It also includes information about the person including name, date of birth, and gender, as well as the position of the face in the image. Gender detection and age estimation are usually performed using this dataset.

computer vision datasets
Figure6. The IMDB-Wiki dataset. Source
Related articles
From NLP to Computer Vision
In this article, we discuss the Attention mechanism and trace its history of use from natural language processing to...
TensorRT
TensorRT is a library developed by NVIDIA for faster inference on NVIDIA graphics processing units (GPUs). It can improve...
Generative Adversarial Network
A GAN (Generative Adversarial Network) is a framework to estimate generative models. This framework includes two models: generators and...
What is a Neural network
This article introduces the concept of Neural Networks in detail. We will compare neurons in human brains with artificial...
Machine Learning
This article will explain Machine Learning concepts in detail through real-world examples. We will then go over various types...
Introduction to Image Segmentation
Image Segmentation is one of the most crucial tasks performed by computer vision models. In this article, we discuss...
Subscribe to our newsletter and get the latest practical content.

You can enter your email address and subscribe to our newsletter and get the latest practical content. You can enter your email address and subscribe to our newsletter.