Ethnic Non-Heritage and Jewellery Industry Cluster Development Strategy Under the Perspective of Big Data

Abstract

In the field of jewellery design, the integration of aesthetics and technology has become the highest significance. Creating exceptional works requires an extensive comprehension of various physical appearance, preferences, and cultural impacts. The incorporation of person body recognition technology is emerging as a revolutionary approach. This novel method ensures customized design solutions, adapting jewellery to individual body types and preferences, introducing a new generation of accuracy and personalization in the dynamic field of jewellery designing. This study  aims to enhance the accuracy and efficacy of person body recognition in the field of jewellery designing model through the implementation of the Social Spider optimized Long Short-Term Memory (SS-LSTM) approach. Initially, we gathered dataset from online open-source data packages, which included various human part images (such as fingers, legs, hands, feet, etc). The gathered data was pre-processed using the data cleaning concept, and we extracted the significant features of the gathered data like feet, hand, and face for train our proposed methodology. The social spider optimization (SSO) algorithm is used to enhance the features of suggested Long Short-Term Memory (LSTM) to develop a novel method for recognizing the person body for a jewellery designing model. We implemented our proposed SS-LSTM in Python 3.10, and recognition accuracy for various body parts like face, feet, hand and fingers was determined. We performed a comparative analysis on various parameters, including accuracy (98.65), recall (98.62), precision (98.67), and F1-score (98.69), assessing our suggested SS-LSTM methodology with existing approaches. The proposed SS-LSTM methodology exhibited superior results in comparison to previous methodologies.