Lubriscan: A deep learning approach for revealing counterfeit motorcycle oil/lubricants via primary packaging analysis

By: Abrera Jr., Joselito Joshua S.; Caballes, Gab P.; Santiago, Denmark O
Language: English Publisher: . . c2025Description: Capstone Project: (Bachelor of Science in Information Technology) - Pamantasan ng Lungsod ng Maynila, 2025Content type: text Media type: unmediated Carrier type: volumeGenre/Form: academic writingDDC classification: . LOC classification: T57.7 A27 2025
Contents:
ABSTRACT: The practice of selling counterfeit motorcycle lubricants have long invaded the Philippine market most notably on e-commerce platforms such as Shopee and Lazada where products cannot be seen and manually verified. Consumers are also buying these products without ever knowing they are counterfeit in the first place due to limited knowledge. To combat this, LubriScan is a mobile application developed to accurately detect and classify the authenticity of motorcycle lubricants through their primary packaging, aiming to enhance user accessibility and awareness towards counterfeit products. The mobile application utilizes the latest YOLOv8 multi-label classification algorithm to identify key features on the bottle of the lubricant and enhance user’s awareness on the lubricants they purchase. The model was trained by purchasing genuine lubricants from authorized resellers while counterfeit lubricants were obtained from Shopee by inspiring the number of 1-star reviews provided by consumers after acquiring these products. 310 images of counterfeit lubricants and 249 images of genuine ones, with a total of 559 images were used as dataset for the model. The model was validated using 112 images from the dataset and achieved a Mean Average Precision (mAP) of 91%, demonstrating high precision and recall. Additionally, the model achieved an accuracy of 88.4% and an F1 score of 90.44% with few instances of false positives and false negatives due to misclassification. Motorcycle riders and motor shop staff and owners evaluated LubriScan using the ISO/IEC 25010 software quality model under the Functional Suitability, Performance Efficiency, Usability, and Reliability. The system received an overall mean score of 4.48 as “Satisfied”, indicating the respondent’s positive reception of the app’s capabilities. LubriScan effectively combines accessibility and accurate detection, making it a valuable tool for both consumers and retailers in spotting counterfeit motorcycle lubricants as well as enhancing awareness towards motorcycle lubricants.
Tags from this library: No tags from this library for this title. Log in to add tags.
    Average rating: 0.0 (0 votes)
Item type Current location Home library Collection Call number Status Date due Barcode Item holds
Thesis/Dissertation PLM
PLM
Filipiniana Section 		Filipiniana-Thesis T57.7 A27 2025 (Browse shelf) Available ft8777
Total holds: 0

ABSTRACT: The practice of selling counterfeit motorcycle lubricants have long invaded the Philippine market most notably on e-commerce platforms such as Shopee and Lazada where products cannot be seen and manually verified. Consumers are also buying these products without ever knowing they are counterfeit in the first place due to limited knowledge. To combat this, LubriScan is a mobile application developed to accurately detect and classify the authenticity of motorcycle lubricants through their primary packaging, aiming to enhance user accessibility and awareness towards counterfeit products. The mobile application utilizes the latest YOLOv8 multi-label classification algorithm to identify key features on the bottle of the lubricant and enhance user’s awareness on the lubricants they purchase. The model was trained by purchasing genuine lubricants from authorized resellers while counterfeit lubricants were obtained from Shopee by inspiring the number of 1-star reviews provided by consumers after acquiring these products. 310 images of counterfeit lubricants and 249 images of genuine ones, with a total of 559 images were used as dataset for the model. The model was validated using 112 images from the dataset and achieved a Mean Average Precision (mAP) of 91%, demonstrating high precision and recall. Additionally, the model achieved an accuracy of 88.4% and an F1 score of 90.44% with few instances of false positives and false negatives due to misclassification. Motorcycle riders and motor shop staff and owners evaluated LubriScan using the ISO/IEC 25010 software quality model under the Functional Suitability, Performance Efficiency, Usability, and Reliability. The system received an overall mean score of 4.48 as “Satisfied”, indicating the respondent’s positive reception of the app’s capabilities. LubriScan effectively combines accessibility and accurate detection, making it a valuable tool for both consumers and retailers in spotting counterfeit motorcycle lubricants as well as enhancing awareness towards motorcycle lubricants.

Filipiniana

There are no comments for this item.

to post a comment.

© Copyright 2024 Phoenix Library Management System - Pinnacle Technologies, Inc. All Rights Reserved.