Multifaceted Rose Leaf Disease Dataset for AI-Driven Plant Pathology

Published: 17 May 2024| Version 1 | DOI: 10.17632/8jtfk9szbg.1
MD Hasan Ahmad


1. Roses are highly valued in the horticultural industry due to their aesthetic appeal and monetary worth. However, numerous diseases that can seriously affect plant health and productivity often pose a challenge to their cultivation. It is essential to accurately and promptly identify leaf diseases in order to treat these issues. A carefully selected collection, the Rose Leaf Disease Dataset is intended to help horticulturists, researchers, and machine learning professionals identify and categorize illnesses that impact rose leaves. This collection of high-resolution photos of rose leaves displays a variety of situations, including diseases and healthy ones. It covers several prevalent rose leaf illnesses, such as Downey mildew and black spots. Extensive annotations are appended to every image, offering discernible information regarding the kind and intensity of the ailment, thus enabling accurate and dependable model training and validation. Furthermore, the information is enhanced with metadata that documents the geographical locations and ambient conditions of the image captured. This contextual data can improve the precision of predictive models and is essential for comprehending the variables influencing the occurrence of disease. 2. These days, deep learning and computer vision techniques hold a lot of promise for handling these kinds of classification and detection tasks. 3. A thorough dataset for Rose Leaf Disease is presented in order to develop deep learning methods. This dataset's classifications were created in cooperation with a subject matter expert from an agricultural institute. 4. A total of 1000 images of downy mildew, black spot, and fresh leaves were gathered from the Savar demonstration location in Dhaka, Bangladesh. Then, in order to enhance the amount of data points, 30,000 augmented images are created from these original photographs by applying techniques including flipping, width shifting, height shifting, brightening, rotating, shearing, and zooming.



Daffodil International University


Computer Vision, Image Processing, Machine Learning, Image Classification, Deep Learning, Image Analysis