Abstract:
By evaluating the trained dataset, a deep learning model produces astonishing results for
image processing. A Deep learning method has been utilized in our research to
automatically identify iris color, and this system has a dataset of 2705 photos from five
different species of iris. To categorize iris color, four Convolutional Neural Network
(CNN) models were used. These models produce more accurate picture classification
results. The image data must be preprocessed before these models can be applied. Data
preparation necessitates the use of certain methods. RGB conversion, filtering, resize &
rescaling, and categorization are the options. Following the application of these methods,
image data is preprocessed and prepared for use in classifier algorithms. The suggested
model has a 90% accuracy rate and was cross-validated using stratified Cross-Validation.
After that, a transfer learning model called "ResNetV2" was used, which resulted in a score of 74.55 % accuracy.
