<![CDATA[This research takes the form of Flower Species Recognition using Convolutional Neural Network (CNN) to optimize the identification of flower types based on color spectrum. The color spectrum of flowers can vary significantly between species and even within a single species. This can pose a challenge in developing a model capable of identifying flower types with high accuracy amidst a wide spectrum of color variations. Selecting an appropriate CNN architecture and optimizing model hyperparameters to achieve optimal performance is a complex process. Careful exploration of various architectures and optimization techniques is necessary to improve the accuracy of flower type identification. The dataset used is collected from various repository sources, comprising images of flowers captured under different lighting conditions, representing diverse color spectra. In this study, data preprocessing stages include color spectrum normalization, feature extraction, and data augmentation to enhance dataset diversity. The CNN model in this research is optimized through network architecture optimization. Model evaluation is performed using standard performance evaluation metrics such as accuracy, precision, and recall. It is expected that this research will yield a CNN model capable of identifying flower types with good accuracy levels, despite facing a wide range of color spectrum variations. This will facilitate the identification and grouping of flower types based on their visual characteristics..]]>
Copyrights © 2024
