Sustainability is one of the key issues in the fashion industry. According to the Institute of Water Report, 2,720 litres of water is used to create one cotton shirt. Moreover, fashion companies are also finding ways to reduce the wastage of fabric that is used for designing each piece of clothing. Here is where Fractal Analytics comes into the picture.

“Computer vision aids in optimising production in the fashion industry by helping to minimise fabric wastage. Designers can analyse the dimensions of a given piece of fabric and determine the most efficient cutting patterns,” said Prosenjit Banerjee, Director of Machine Vision Research & Practice, AI@Scale at Fractal, in an exclusive interaction with AIM. 

“Though computer vision does not directly reduce the wastage but it is fetching the data that is training the models, which adds to the minimisation that can be done by machine learning models,” said Prosenjit. “A certain size is given, let’s say XL. So given this piece of cloth, how do you cut it so that you know the wastage? Computer vision can actually understand the dimension of the cloth and decide the best possible cutting line so that the cutting process gets automated, and there is less wastage in the end.”

“Computer vision-based algorithms are at the heart of these innovations”

“To understand how computer vision helps in the fashion field, we will have to understand the value chain of the industry,” starts Prosenjit. The fashion value chain begins with sourcing, where raw materials are cultivated and prepared for production. From fibre cultivation to yarn preparation and dyeing, this phase is essential in setting the foundation for quality apparel. 

Prosenjit explains that computer vision has a significant impact in quality control. “Textile mills often employ cameras along the production lines to detect defects and blemishes in the fabric. These cameras are equipped with advanced computer vision algorithms that can identify imperfections. In case of continuous defects, the system triggers an alarm, preventing substandard products from reaching the market.”

Due to this, a lot of waste products are stopped before production, resulting in a reduction of wastage. 

Then comes the marketing phase, where computer vision has a vital role, majorly in trend analysis and prediction. Social media platforms like Instagram and Facebook provide a rich source of fashion-related images and data. By scraping this data and analysing the text and images, computer vision algorithms can identify emerging fashion trends.

Prosenjit explains that this information is valuable for demand forecasting. For instance, during festivals like Diwali, understanding popular clothing trends can help retailers make informed decisions about what to stock. This data-driven approach enables businesses to align their inventory with consumer preferences and reduce the risk of overstocking or understocking, a crucial consideration given the narrow sales windows in the fashion industry.

Leading the sustainability chain

Fractal, the company that was named the leader in computer vision consulting in 2020, is moving beyond just computer vision. It has its machine vision accelerator called IVA, which stands for Image & Video Analytics. IVA contains different models and various API calls which can be used for different functionalities, and not just fashion, from retail to manufacturing, to satellite imagery. 

Fractal aims to build a spin-off called IVA Fashion, which will be especially focusing on the fashion industry. Currently, IVA is working with the Cerebral team on a concept known as Micro Stimuli. The concept is about triggering emotions within customers by targeting specific neurons in the brain, which is still under development. 

We have already heard of virtual try-on, where you can leverage AR/VR for testing out clothes. This involves digitising the human body using 3D modelling, creating an exact replica of an individual’s measurements. By inputting clothing dimensions or 3D models from renowned brands, customers can virtually try on clothes, reducing the likelihood of returns due to poor fit. This technology aims to minimise clothing wastage, thus resulting in overall sustainable goals. 

The post How Fractal is Saving Thousands of Litres of Water Using Computer Vision appeared first on Analytics India Magazine.