BFCAI_project_Flowers_Recognition

AI project using deep learning (CNN) to flower classification.

Links

• kaggel dataset link : https://www.kaggle.com/alxmamaev/flowers-recognition

• kaggel kernal link : https://www.kaggle.com/mohamed6198/cnn-keras-flowers-recognition

Objective

• The goal is to pick an image of flowers then classify it.

• The data files contain 4323 RGB image.

• The data contain 5 types (daisy, Rose, dandelion, sunflower, tulip).

• Each type has its folder contain its images.

Steps

• Loading libraries will be used like (keras, seaborn, pandas, numpy, … ).

• Loading data and discover how it was stored.

• Detecting shape of images to get a shape which the images will be converted to then resize images.

• Normalize all data.

• Splitting data to training part and test part by ratio 9:1.

• Creating model and train it 70 epochs.

• Evaluating mode and read a final value for accuracy.

• Test model in some samples.

Evaluation

• plot graph for train accuracy evaluation with each epoch.

• plot graph for train loss value evaluation with each epoch.

• plot graph for validation accuracy evaluation with each epoch.

• plot graph for validation loss value evaluation with each epoch.

• clasification report.

Output