Executive Summary : | Leather, a durable material made from animal hide, is a widely traded material globally. It is recognized under the 'Make in India' Foreign Trade Policy and the 'Skill India' and 'Digital India' programs. The quality of leather products depends on various factors, including species, breed, age, sex, origin, skin area, leather type, and genuine leather. The leather sector faces a recession due to the sale of non-leather and low-quality animal skin under the genuine leather tag. Leather authentication and species identification are crucial for protecting biodiversity, detecting non-leather, maintaining the quality and value of leather products, consumer protection, and dispute settlement. Traditional techniques like microscopic examination, DNA analysis, and liquid chromatography/mass spectrometry are theoretical, subjective, slow, and supervised. Digital microscopic image analysis using machine vision is more advantageous in terms of optimization, automation, and cost-efficiency. This research aims to develop a reliable, scalable, and flexible digital ecosystem for leather authentication and species identification. The project acquires a dataset of nearly thirty-thousand microscopic leather images of various species, develops efficient deep learning models for authentication and species identification, and integrates image processing, computer vision, and cloud computing with an API for efficient real-time processing of complex image data. The proposed ecosystem contributes to the digital knowledge of animal species in the leather sector, protecting biodiversity and providing a strong platform for digital identity of leather processed from permissible species for consumer protection. |