Introduction

A coin is a small, flat, (usually, depending on the country or value) round piece of metal or plastic used primarily as a medium of exchange or legal tender. They are standardized in weight, and produced in large quantities at a mint in order to facilitate trade. They are most often issued by a government. Coins often have images, numerals, or text on them. Obverse and its opposite, reverse, refer to the two flat faces of coins and medals. In this usage, obverse means the front face of the object and reverse means the back face. The obverse of a coin is commonly called heads because it often depicts the head of a prominent person, and the reverse tails. Coins are usually made of metal or an alloy, or sometimes of man-made materials. They are usually disc shaped. Coins made of valuable metal are stored in large quantities as bullion coins. Reference - Coins

Content

This dataset contains the coin images of various currencies, as listed below:

1. US coins
2. Chinese coins
3. Yen coins
4. Euro coins
5. Indian Rupee coins
6. Peso coins

The coin images are collected by scraped from Google Image searches and filtered, based on the coin image quality. Finally, the number of coins in the images are manually counted, to generate the ground truth.

The coins_count_values.csv file contains the ground truth

Challenges

This dataset can be a good challenge for Kagglers, who are new to Image preprocessing, and also helpful for those who want to practice image preprocessing techniques.

1. All coins are not round shaped
2. Background of the coins is not the same
3. Color of the coins is not the same
4. Most of the coins present in the image collections are filtered based on the position of coins (horizontal flat) for ease of detection, but some coins images are added to challenge your preprocessing algorithm as shown below:

Hope you will find the dataset useful for practicing image processing/ computer vision skills.

Have fun and Happy Learning!