Identification of Stain on White Fabric and Data Set Generation

This paper is about a novel technology, which can be used in washing machines for determination of the washing courses. This work utilizes two efficient deep learning algorithms to identify stains on a fabric. This paper discusses the process of synthetic data-set generation, through which the image data, required for the identification of stains, is generated. The idea is to implement the process on a low-end device, therefore Raspberry PI 3 B+ model was used. The Neural Network models used for the identification of stain were Faster RCNN Inception V2 Pets and SSDlite MobileNet V2 COCO. This module can be used to identify the stain on fabric, that would allow pre-processing on fabric and determine washing courses.