Business Goal
People hestitate when it comes to purchasing clothes online. Part of the reason lies in not knowing the size or fit. To address this issues we propose a system which helps recommend your T-shirt size.
- Allow system to determine T-Shirt size using webcam
- Allow system to disaply filtered results accordingly
- Support saving size information and account into different brand sizes and measures
A Mockup of how the user would interact with our system
Report
In this paper, we demonstrate how image analysis can be used to implement this feature. To understand its impact, we also conducted an analytical and empirical evaluation of our system.
The following diagram depicts our approach to this project
The low-fidelity prototypes helped us quickly reach a good design. Empirical evaluation in the form of user feedback (before and after use of application) helped us in determine improvements in the systems' interaction.
Running the Code
To run the code, simply clone the repository and inside Code double click the home.html. The backend service is hosted live, so the image is sent and the system recieves the size prediction response. Just follow the steps described in here
Conclusion
While the system promises better interaction while shopping on online e-commerce systems, better image analysis could help overcome using of object (CD in this case). Many companies including Lenskart and RayBan have launched similar features in their system to help users try glasses online.
Attribution
This project was designed and developed by Shrenuj Gandhi, Chiraj Kapadia, Shristi Arora and Adtiya Mahandhare under the supervision of Dr. Robert St. Amant.