RosePixels- The Dutch Rose Petals Classification Dataset

Published: 20 May 2025| Version 3 | DOI: 10.17632/y53kbt9jss.3
Contributors:
Arya Sabale, Devashri Kapse, Mahek Mehta,
, Pranali Mrs.Pranali Gajanan Chavhan, Priti Rathod, Gagandeep Kaur

Description

The RosePixels dataset is a detailed collection of high-quality images featuring Dutch rose petals, gathered straight from the Talegaon-Maval region—a place famous for growing these beautiful flowers. Unlike other datasets that tend to focus on just a few aspects, this one offers a broader view, capturing a wide range of petal colors and their health conditions. The dataset comprises images of petals in shades like red, yellow, pink, purple, white, orange, plus stunning bi-color and multi-color varieties. On top of that, it sorts them by health—whether they’re fresh, dry, or diseased. In total, RosePixels includes 1,906 images. Breaking it down by color, there are 298 red, 210 yellow, 218 white, 166 pink, 220 purple, 148 orange, 185 bi-color, and 135 multi-color petals. As for health, 913 are fresh, 667 are dry, and 326 are diseased. All these images were captured in a controlled setting, and each one has been pre-processed with a unique ID—like "FRESH_RED_001," where "FRESH" shows the petal’s condition, "RED" is the color, and "001" is a three-digit code to keep everything organized. This dataset is perfect for training machine learning models, especially for quality checks or projects tied to agriculture, perfumes, edible products, flavors, or even cosmetics. It’s a treasure trove for anyone looking to explore the world of Dutch roses!

Files

Institutions

  • Vishwakarma Institute of Information Technology

Categories

Artificial Intelligence, Computer Vision, Plant Biology, Machine Learning, Rose, Convolutional Neural Network, Deep Learning, Agriculture

Licence