Why did we create this application?
Are sneakers in fashion? That's for sure.
Everyone talks about sneakers, buys sneakers, wears sneakers, collects sneakers, sells sneakers...
The sneaker market represents a significant part of the global economy.
Enough about economics.
Have you ever seen people wearing models you love without knowing the name of the model or where to find it?
The new Findout Sneakers® concept aims to allow its users to take a picture of a pair that he likes in order to be redirected to a platform with all the necessary information and links to buy it!
Scroll down in order to know more about the process making it works!
It will consist in classifying a huge number of sneakers in the data base. Diverse criteria will be assigned to each model of sneaker classified such as:
- Shape
- Logo
- Colors
These criteria will classified the models according to their names, the date of release and links to the website to buy it.
In order to feed that data base, we will use the "ImageClassifier" function with ml5.js.
This function will consist in using neural networks: using Artificial Intelligence to recognize the content of the photos and pictures people will upload on the application.
Then, the pictures will be classified and associated with a particular model of sneakers. However, in order to allow a precise identification of the model photographed we must provide the database with a huge number of pictures of sneakers taken in order to train the neural network and to make this Artificial Intelligence process used to it.
Therefore, different pictures of sneakers will have to be taken and added to the AI. Do not hesitate to use the photos taken by users to add them to the database. Indeed, the more photos will be added to the database, the more efficient the process of recognizing the sneakers models will be. Indeed, the database will be more important in terms of quantity and quality.
Nike
Example 1
Nike Air Pegasus 92
October 2013
New Balance
Example 2
New Balance 630 Brorange
May 2017
Adidas
Example 3
Adidas Original Spezials
July 2013
Reebok
Example 4
Reebok CN 7119
July 2014
Concerning the process of recognizing the elements present on the image in order to associate the photo with a particular sneakers model, here is how it works:
Example: This picture was taken on a stranger in a bar on November 25th 2019 in Lyon, France.
1st - Focus recognition
2nd - Shape recognition
3rd - Logo recognition
4th - Colors recognition
After these four steps performed by the recognition algorithm, the sneaker model photographed is associated with the corresponding model stored in the database.
After processing, the result of the artificial intelligence is as follows: associated model: "New Balance U420 Burgundy".
The developer job here consists in creating an intuitive page for the user where he can find: an original photo of the product, its technical elements such as the destination use of the model, for which season as well as its different composition elements (materials used, colours).
Finally, there should also be the price displayed as well as a possibility of purchase through links to various sites with which the application has developed partnerships such as FootLocker, Courir, Flightclub, Hanon and many others.
Let's have a look on how it can be designed with 2 Figma pictures you can also find here:
Findout Sneakers - DWW Project on Figma
Picture page
Info and basket
Check here two possible shopping page templates found on Codepen.io for the products you can develop in coding with HTML, CSS and JS:
First example
or
Second example
All the information you will need in order to develop the app