provides a rich shopping experience for the users of Amazon by providing Augmented environment where customers can visualize a product that will be involved in Machine Learning pattern recognition and recommender system for best exposure and maximum satisfiabilty to the customer with no need left to look around for better options!
The Augine is a multi-platform 3-dimensional shopping experience providing algorithms that makes it possible to recommend customers with similar products based upon his preferences which can then be used for a virtual 3D try-on using concepts of Computer Vision and Augmentation Reality.
To create an AR appplication in interior design which gives an ability to the user to create a space the way they want to. This AR visualization will help in accelerating the purchases of customers and to decrease the possibility of return of the products which are cancelled due to undesirable conditions. There is a severe need to provide tools which help interior designers.
There is a majority of customers who find out that the clothes they purchased are not fit to their size or that the texture of the cloth is not what they expected. Providing a real-time dressing room environment to test out the product without worrying to make the order will help boost the efficiency and success rate of every transaction on Amazon.
The complete data beginning from preferences and ending to successful orders when stored in a database makes a solid ground for customers to keep using Augine for Amazon in many upcoming years.
-
It will be a revolutionary solution to the constraints faced by customers and retailers of Amazon in visualizing the actual product even before it is created which saves time and resources.
-
Providing a Javascript interface backed by Django to maintain a SQL database provides a 3D try-on when integrated with MLP(which is basically a feedforward artificial neural network). User Interface
-
Machine learning will recognize the operational patterns of the preferences of the customer based upon his fondness of products as well as the attributes we will pick from his environment and develop upon them to predict and recommend new products which will cater his needs to the maximum possible extent.
-
With our Neural networks model, we will be able to synthesize personalized suggestions picking up from the data stored in our dynamic database to save the history as we go for a long term solution.
-
User will also get an ability to edit the designs and make changes even if the design is at the final stage. So, now users need not worry about the tedious corrections that are done in actual décor and furniture.
-
AR makes it possible to measure any dimension with the help of a smartphone's camera. The customer can aim/target it on a detected plain
-
AR in interior design will let a user guide the designer in the best possible manner. Even the minute details related to the designing process can be communicated interactively using AR.
-
It can either take a real garment as input, or generate it from scratch, and drape it on top of the SMPL body for any shape and pose.
Currently, every available AR headset is a bulky piece of hardware that may be too expensive for the masses, making the entire experience limited and inconvenient.
Mobile AR faces major issues in displaying visuals accurately. Additionally, smartphone cameras are built for 2D image capture and are incapable of rendering 3D images. Hence, the hardware required for AR technology needs to be enhanced before mass adoption.
One would need about a million relevant records to train an ML model on top of the data. And it cannot be just any data. Data feasibility and predictability risks jump into the picture.
Assessing if we have relevant data sets and do we get them fast enough to do predictions on top isn’t straightforward. Getting contextual data is also a problem.
Testing machine learning models is difficult but is as important, if not more, as other steps of the production process. Understanding results, running health checks, monitoring model performance, watching out for data anomalies, and retraining the model together close the entire productionizing cycle.
The dependencies and the terminals required.
Want to contribute? Great!
For production release:
127.0.0.1:8000