try on the go artificial intelligence and computer vision powered virtual try on for e commerce11736

Project TitleProject Area of Specialization Artificial IntelligenceProject Summary

“Try on the go” will provide people with an online alternative to an offline store’s try room where you can try on different outfits to see if they are the right match using your web camera. The motivation behind this project is to minimize the number of returns people make when they end up with clothes that don’t look good on them; by providing them with a virtual clothing system.

Project Objectives

The objectives of this project are:

• A virtual try-on system.
• Minimizing the number of returns people make when they end up with clothes that don’t look good on them.
• Reducing the amount of time people spend in offline store’s try rooms.
• Building trust.
• Improving customer retention.

Project Implementation Method

In the proposed work, our main goal is to implement a practical solution that can enhance customer satisfaction by providing an augmented reality experience through artificial intelligence and computer vision. The proposed system will be developed with a focus on minimizing cost and hardware components. The system will be more accessible to customers and will make use of mobile/computer web cameras to capture human motion and virtually fit clothes on them.

This application will allow users to:

• Select/Upload a picture of the item they like,
• Turn on their web camera and either capture a photo or import it from the gallery,
• Try the outfit on themselves virtually to see if it looks good on them without the need of spending hours in an offline store’s changing room.

This can be done by making an application where users will select or upload the picture of the outfit they like, click on the try-on button, either open their cameras and take a photo or upload it from their gallery into our application.

The system will find the key points on the human body and the outfit they want to try using OpenPose and OpenCV libraries in Python. The coordinates will be matched and the desired outfit will be overlaid on the uploaded picture of the user using computer vision. The application will be developed using Python and its frameworks and libraries. User’s photos will be secured using some encryption and access control mechanisms.

Benefits of the Project

The average cloth return rate rests at around 20% while brick-and-mortar return rates averages at 8%. Most apparel companies have to deal with high return rates because of customers’ inability to try on the clothes physically before buying them as they can do in offline stores.

At present, people are living a busy lifestyle which is leading them to buy ready-to-wear clothes with or without try-on, hoping for a perfect match. Trying on clothes physically is a difficult and time-consuming process to find the best matching outfit. Also on the vendor’s side, it is a tedious process to repack all the returned or misplaced items that were taken out by the customer.

In online shopping, it is virtually impossible to find a suitable outfit that would suit them by just looking at a few images of clothes. People are unable to see if the outfit they are buying through an online shopping platform will be a perfect match. Hence, this leads to several issues like the customer is not satisfied with the product and now he wants to return it.

To overcome this problem, virtual dressing rooms come into play where it allows customers to try on different apparel virtually in front of their web cameras to choose the best-matched outfit within a shorter time from the comfort of their homes without the need of spending hours in an offline store and enhances customer satisfaction.

“Try on the go” will help both customers and online shopping platforms by allowing customers to select the best outfits for themselves by trying them virtually and thus reducing the return rates and leveraging customers’ curiosity about the product and intention towards purchasing the product.

Technical Details of Final Deliverable

The expected outcomes of the project are:

• A module that makes online shopping more practical and usable than it is now. An experience that makes you feel as close to physically shopping while shopping online. With easy access through a cell phone camera or a webcam, a user will feel more comfortable selecting options for their wardrobe online.
• The module will be fast and efficient so that the process is as quick as possible.
• Fitting of clothes from the user’s uploaded cloth images on the human body image(s).
• Processing time to be under a few seconds.

Final Deliverable of the Project Software SystemCore Industry ITOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Industry, Innovation and InfrastructureRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	80000
GPU	Equipment	1	70000	70000
Datasets	Miscellaneous	1	10000	10000