Inspiration

Fashion Retailers try to draw inspiration from external sources such as e-commerce portals and online fashion magazines to design the next set of fashion products that they can launch in order to delight the customer. However, it is a manual effort-intensive process, requiring a large team of fashion designers. In order to reduce dependency and make the overall process more efficient, the Retailer wants a  scalable tech solution to extract winning designs of apparels and footwear in a consumable format so that the same can be incorporated in the upcoming design of fashion products.      Problem Statement  A fashion retailer wants to source ongoing and upcoming fashion trends from major online fashion portals and online magazines in a consumable and actionable format so that they are able to effectively and efficiently design an upcoming fashion product portfolio. 

What it does

A one-Stop platform for fashion designers which scrapes the internet for the trending fashion websites and presents it in a single and easily consumable format. It is also capable of giving sales recommendations to retailers and tell which designs have a high sale probability using the power of convolutional neural networks. This tool is also smart enough to auto-generate new design patterns from existing ones using a synchronous, interactive, generative, adversarial network. The web sections include the main page where all the sites state who is presented in views. Here you can find the trending designs of the top fashion Web sites on the Internet. The next section is the celebrity section. This section keeps track of the brands endorsed by celebrities and the designs they wear. Retailers can keep track of these designs and implement them in their manufactured products. Next comes the sales recommendation section. This section shows how well a particular design would do in the market. It takes into consideration previous reviews of the customers and the type of designs they're like. Let's talk about the future scope of this platform. One would be the addition of virtual try-on 3D human models to enhance designers experience and improve on designs using augmented reality next to a ladder users administration portal to classify and give access to users and manage their feed based on interests. Panels for marketing team sales teams and design team to virtually collaborate and share designs and inspirations virtually. We would also integrate designing and editing tools in the solution so as to establish it as one-stop for all needs of designers and teams.

How we built it

The project is built over the strong base of Atlassian connect and followed by a backend using scrapy , AI GAN models in the following ways:-

  • Scrapy to scrape up to date fashion trends from across the internet in form of link and images to the apparel.
  • Cloud service (FAAS) on Atlassian Connect to enhance the scalability of the product for large scale user group
  • AI GAN Model to generate new designs from already existing trendy designs.
  • Powered with CNNs to give sales recommendation to increase sales probability( By using audience and retailer reviews it learns peculiar details from the data)
  • High CI/CD support as Project built as a confluence macro hence deployment within seconds .
  • Cross Platform cloud support to ensure maximum reach to users across devices and platforms.

Challenges we ran into

It was a complete thrill-packed experience, getting started with Atlassian cloud products seemed like a difficult step to do, but thanks to the great documentation and video tutorials provide at Atlassian site, made it extremely easy for us to understand the workflow to make cross-platform apps quite easily

Accomplishments that we're proud of

We are proud to build upon technologies which were more or less alienated from us and worked under a short window of time to give life to this idea of ours by building an industry-ready platform from scratch.

What we learned

We learned to use the atlassian connect platform and to integrate scrapy and other such libraries. We also learned about DL model and it's realtime applications.

What's next for FashionKart

  • Addition of virtual try-on on 3d human models to enhance designers experience and improve on designs and add AR try on features.
  • Add a users portal to classify and give access to users manage their feed based on interests.
  • Adding panels for marketing team , Sales teams and design team to virtually collaborate and share designs and inspirations virtually.
  • Integration of designing and editing tools in the Solution so as to establish it as one stop for all needs of designers and teams.

Built With

Share this project:

Updates