Textile defect images defection through deep learning approach

Project Title

Textile defect images defection through deep learning approach

Project Area of Specialization Artificial IntelligenceProject Summary

In modern textile industrial processes,textile defect recognition is a very important task for quality control of this vital industry. As deep learning has revolutionized the machine learning domain from classification to regression.This project will evalaute the potentialso of critical machine-learning method, the deep convolutional  neural network (CNN) , for recognition of the textile defect images.

Project Objectives

to do the singel label classification of fabric defect images using deep convolutional neural network and comparing its results with other deep learning methods.

To validate the proposed methodology

Project Implementation Method

Tools:

Python, Image preprocessing And Convolutional Neural Network, Tensorflow and Keras 

Methodology:

Import libraries

Collecting and loading Textile defect images

Preprocessing 

Building CNN

Training the CNN

Predicting

Evalaute

Benefits of the Project

This project will be able to classsify fabric defect using deep learning reducing the error rate in human inspection which is 40% to 60%.

Less defect textile results in more economic and profit growth as the indsutry faces much of the profit loss due to poor or defected quality of fabric.

Technical Details of Final Deliverable

High Processing Power

Increased Ram for training the model and preproccesing of the images

Final Deliverable of the Project Software SystemCore Industry ManufacturingOther Industries Others Core Technology Artificial Intelligence(AI)Other Technologies OthersSustainable Development Goals Decent Work and Economic Growth, Industry, Innovation and InfrastructureRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	42000
Google Colab Pro+	Equipment	1	10000	10000
2 GB Grahpic card	Equipment	1	22000	22000
Keyboard	Miscellaneous	2	3000	6000
Mouse	Miscellaneous	2	2000	4000