MAC FRUIT CLASSIFIER

STACK

Machine Learning

PROJECT

Chakka vs Manga Image Classifier It is a fundamental Deep Learning project done using Python Keras. In this Keras project, we will discover how to build and train a convolution neural network for classifying images of Jackfruit and Mango. The dataset includes 500 images with equal numbers of labels for jackfruit and mango.

LINK TO VIDEO:

https://drive.google.com/file/d/1B8myR7rcM9viBfHuz9VOCAoDfe2nG-LX/view?usp=sharing

TEAM ID

BFH/recMuXbQkYDR42dsR/2021

TEAM MEMBERS:

1. ANJALI RAJENDRAN

2. ANJALY SATHEESH

3.ALENA ACHANKUNJU DANIEL

How it Works?

1.Import the libraries

2.Define image properties

3.Prepare dataset for training model

4.Create the neural net model

5.Define callbacks and learning rate

6.Training and validation data generator

7.Model Training

8.Save the model

9.Test data preparation

10.Make categorical prediction

11.Convert labels to categories

12.Visualize the prediction results

13.Test your model performance on custom data

14.To install flask

15.Now create a new directory, copy your model (“model.h5”) to this directory

16.Create a file app.py and input codes, make a webpage body design using html

17.Save this file and run using python conda venv

How to configure

1.use Python (3.x)

2.import required libraries

3.Dataset should be divided into two folder: Training and Testing

4.Estimate models perfomance and validation accuracy; the more it is accurate, the more it will predict the image correctly

Libraries used

1.tensorflow==2.0.0

2.tensorflow-estimator==2.0.1

3.Keras==2.3.1

4.Keras-Applications==1.0.8

5.Keras-Preprocessing==1.1.0

6.Flask==1.1.1

7.importlib-metadata==1.2.0

8.ipykernel==5.1.3

9.Jinja2==2.10.3

10.json5==0.8.5

11.matplotlib==3.1.2

12.notebook==6.0.3

13.numpy==1.17.4

14.opencv-contrib-python==3.4.0.12

15.opencv-python==4.1.2.30

16.seaborn==0.9.0

17.sklearn==0.0

18.Werkzeug==0.16.0