Date fruits classification using AI

Project Title

Date fruits classification using AI

Project Area of Specialization Artificial IntelligenceProject Summary

Date fruits are small fruits that are abundant and popular in the Pakistan, and have growing international presence. There are many different types of dates, each with different features. Sorting of dates is a key process in the date industry, and can be a tedious job. The aim of this project is to classify the types of date fruit using machine learning methods.

Project Objectives

There are over 40 unique types of dates, and over 400 varieties, which cover a wide range of taste, shape, and color, as well as price and importance. The process of classifying dates is thus quite important, particularly because a large percentage of consumers can not differentiate between many different types – and thus,

1- One could even envision a cell phone camera-based application, which could be used by consumers in the marketplace to identify dates after capturing its image.

2- It is also particularly important to be able to classify dates visually for automated factory classification.

Project Implementation Method

1. Data Collection and Pre-processing:

Firstly, we will collect the images of 10 types of Date fruits for dataset. We will use mobile camera to capture images for our model. We will be using white background during image Capturing. Then we will remove that background openCV library to prepare dataset for the model. We will collect 200 images for each type. The dataset will be splitted into train, validation and test set (60,20 ,20)% respectively.

For feature selection, we will be using Contrast, Correlation, Energy, Entropy, homogeneity, Eccentricity, RGB, Area, Perimeter, Roundness.

1. Model Selection and Training:

After the dataset is collected, next step is to select appropriate model for training classification of date fruits. We will use LDA (Linear Discriminant Analysis)  Support Vector Machines(SVM) and CNN (Convolutional Neural Network) models. In the first approach, LDA will work as a classifier and posteriorly it will reduce the dimensionality of the dataset and a neural network will perform the classification task, the results of these three  approaches will be compared afterwards.

LDA:

LDA is a supervised learning algorithm that finds the linear combination based on different features that can split two or more classes.

CNN:

In deep learning, a convolutional neural network (CNN/ConvNet) is a class of deep neural networks, most commonly applied to analyze visual imagery. Now when we think of a neural network we think about matrix multiplications but that is not the case with ConvNet. It uses a special technique called Convolution. Now in mathematics convolution is a mathematical operation on two functions that produces a third function that expresses how the shape of one is modified by the other.

SVM:

Support Vector Machine” (SVM) is a supervised machine learning algorithm that can be used for both classification or regression challenges. However,  it is mostly used in classification problems. In the SVM algorithm, we plot each data item as a point in n-dimensional space (where n is a number of features you have) with the value of each feature being the value of a particular coordinate. Then, we perform classification by finding the hyper-plane that differentiates the two classes very well .

1. Performance Analysis:

Finally, we will check for accuracy among different model's results, and compare the accuracy of different models using confusion matrix and ROC curve.

Confusion Matrix:

A confusion matrix is a performance measurement technique for Machine learning classification. It is a kind of table which helps you to know the performance of the classification model on a set of test data for that the true values are known.

ROC Curve:

A receiver operating characteristic curve, or ROC curve, is a graphical plot that illustrates the diagnostic ability of a binary classifier system as its discrimination threshold is varied.

Benefits of the Project

There are over 40 unique types of dates, and over 400 varieties, which cover a wide range of taste, shape, and color, as well as price and importance. The process of classifying dates is thus quite important, particularly because a large percentage of consumers can not differentiate between many different types – and thus,

1- One could even envision a cell phone camera-based application, which could be used by consumers in the marketplace.

2- It is also particularly important to be able to classify dates visually for automated factory classification.

Technical Details of Final Deliverable

This research can be used to provide consumers with a date fruit classification programme using a mobile phone software. Consumers are expected to be able to obtain information on the type of date fruit offered over the counter using the created smartphone application. It is considered that by extracting more features in classification study, success rates in the categorization of not just date fruits but also other vegetables, fruits, legumes, or any object can be improved.

Final Deliverable of the Project Software SystemCore Industry AgricultureOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Decent Work and Economic Growth, Industry, Innovation and Infrastructure, Responsible Consumption and ProductionRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	15000
Dates	Miscellaneous	10	1000	10000
Moblle Stand	Equipment	1	5000	5000
Travel for Research	Equipment	0	0	0