Artificial Intelligence in Marketing Practices: Gait Based Age Estimation using CNNs

Abstract

Every day, new developments in technology significantly influence business activities, especially in marketing. This study explores the application of artificial intelligence to marketing by evaluating how different Convolutional Neural Network (CNN) architectures perform in gait-based age estimation, a key biometric feature for personalized marketing strategies, such as targeted advertising and real-time consumer profiling. Using Gait Energy Image (GEI) representations and the OULP-Age dataset, we benchmarked models including EfficientNetB7, RegNetY, CrossVIT, InceptionV3, and ResNet152. Among these, CrossVIT achieved superior performance with a Mean Absolute Error (MAE) of 2.69 years, outperforming all previously published methods, which typically report MAEs in the 3.7–8.4 range. These findings demonstrate the potential of gait-based deep learning approaches not only for advancing intelligent, real-time marketing solutions but also for contributing to the broader field of computer vision, illustrating how advanced neural architectures can effectively address complex behavioral biometric tasks in real-world scenarios.