Development of AI-based software of mango variety identification

Mango is a rich source of nutrients fulfilling the dietary requirements of consumers. Mango fruit contains an array of biochemicals having anti-oxidant activity. Chromatographic and spectroscopic techniques are used for the isolation and identification of mango biochemicals. Different

Project Title

Development of AI-based software of mango variety identification

Project Area of Specialization

Artificial Intelligence

Project Summary

Mango is a rich source of nutrients fulfilling the dietary requirements of consumers. Mango fruit contains an array of biochemicals having anti-oxidant activity. Chromatographic and spectroscopic techniques are used for the isolation and identification of mango biochemicals. Different mango varieties are shown to have characteristic variations in their HPLC chromatograms based on biochemical components. By exploiting the variation in mango fruit biochemical content, we aimed to develop an AI based software for mango variety identification. Currently, there is no such in-silico technology available to validate claims of the food industry or to evaluate mango cultivars. Utilizing the variety-specific data of mango biochemical contents, we wish to establish characteristic spectral peak patterns of Pakistan’s commercial mango varieties and then applying tools of artificial intelligence to device software. Such an efficient identification program would prove to be helpful in the quality characterization of mango cultivars and post-harvest handling optimizations to conserve the fruit's antioxidant capability.

Project Objectives

• To develop software for mango variety identification based on HPLC patterns.
• To contribute and increase the spectrum of research done in the arena of mango phytochemical analysis.
• Gaining novel insights into Pakistan’s mango varieties
• Exploring content uniqueness of Pakistan’s mango varieties
• Efficient identification of mango varieties in an economical way
• Eliminating the necessity of rigorous and technically demanding spectroscopic analytical procedures to analyze mango phytochemical.
• To bring a change in trends of mango-based products in the market. 
• To promote the consumption of mango varieties having high polyphenolic contents and thus more potential benefits.
• To highlight and prevent adulteration in mango based food products.

Project Implementation Method

1. Mango fruit pulp shall be subjected to biochemical extraction through maceration with the fixed combination of polar solvents.
2. Fruit pulp extracts shall then be centrifuged and evaporated under pressure to separate unwanted carbohydrates and other polar solvents.
3. Aqueous extracts shall then be utilized for HPLC analysis to establish variety-specific patterns.
4. To classify the mango varieties we will use machine learning algorithms. We will test several algorithms such as support vector machines, decision trees, random forests, and logistic regressions.
5. We will perform a comparative analysis to find out the best algorithm among these.
6. Scikit learn (python implementation of machine learning) shall be utilized to extract information and identify patterns in chromatographic data.
7. Train data sets and validate them under supervised learning
8. The established method shall then be applied to sample data sets to obtain identification results.  

Benefits of the Project

The present project is anticipated to provide the following benefits;

• Provide novel and efficient means of mango variety identification
• Evaluating mango cultivars and speculating antioxidant capabilities based on biochemical content of each variety.
• Promoting the use of mango variety, offering maximum health benefits
•  Providing recommendations to farmers for enhancing mango plant polyphenolic content
• Helping farmers to optimize post-harvest conditions to control loss of fruit phenolic content.
• Assisting food industry to validate their claims of providing 100% natural mango based edible products

Technical Details of Final Deliverable

• A software for the identification of specific mango variety will the final product of the present project.
• This AI-based software shall enable users to input structured information regarding sample chromatographic patterns and compare it with the built-in database to identify the variety.
• The software shall also provide supplementary benefits of expeditious results.
• Shall provide additional information on sample data polyphenolic content to estimate the biological value of variety.
• Software shall help to predict climatic and post-harvest conditions contributing to decreasing in antioxidant capacity of sample mango fruit.  

Final Deliverable of the Project

Software System

Core Industry

Agriculture

Other Industries

Food

Core Technology

Artificial Intelligence(AI)

Other Technologies

Sustainable Development Goals

Good Health and Well-Being for People

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com