Mangifera2012: An Image Dataset of Various Bangladeshi Mangoes

Published: 14 August 2024| Version 1 | DOI: 10.17632/w5jg84txj8.1
Contributors:
, Md Masudul Islam,

Description

A comprehensive image dataset of various types of reknown mangoes in Bangladesh. This dataset contains a total of 2012 HD image data in JPEG format. The image resolution is 3024x4032 pixel. The mango variety is collected from different marketplace of Dhaka during Mango season usually from May to August. There are 10 types of mangoes in this dataset which contains un-uniforme numbers of image data: 1. Amrapali (252 images) 2. Bari-4 (235 images) 3. Bari-7 (176 images) 4. Fazlee (156 images) 5. Harivanga (202 images) 6. Kanchon Langra (210 images) 7. Katimon (163 images) 8. Langra (202 images) 9. Mollika (221 images) 10. Nilambori (195 images) Each of the variety images are unique and captured from various angle so that the manoges' size, shape and texture can be recognize shaprly during image processing. Also, the Dataset is in ZIP format where it contains root folder "Mango_Dataset_2012" which contains 10 sub folders for each variety of mango. Moreover, this dataset is a state-of-the-art mango variety dataset and one of the largest mango variety dataset.

Files

Institutions

Bangladesh University of Business and Technology

Categories

Image Processing, Machine Learning, Fruit, Image Classification, Mango, Convolutional Neural Network, Deep Learning, Agriculture

Licence