Akademik Veri Yönetim Sistemi
9th International Congress on Information and Communication Technology, ICICT 2024, London, İngiltere, 19 - 22 Şubat 2024, cilt.1002 LNNS, ss.237-247
The importance of deep learning applications in our daily lives is growing in an increasingly complex world. Large amounts of data can be analyzed using deep learning, allowing for the identification of patterns, which can aid in solving problems in healthcare, finance, and agriculture. It can automatically extract the features of images captured for many different purposes. This study aims to detect apple leaf diseases using Vision Transformers from image datasets captured under contrasting conditions and compare their performance. Moreover, a brief overview of image processing and computer vision algorithms and applications is discussed. The results of our study indicate that deep learning models outperform traditional methods often used in image processing approaches in terms of accuracy. They also improve test accuracy compared with other studies in the same area of plant diseases. Furthermore, we conclude that the conditions under which images are acquired influence the performance of classification models. Thus, choosing the appropriate acquisition conditions requires careful consideration.