The Influence of Visual Features in Product Images on Sales Volume: A Machine Learning Approach to Extract Color and Deep Learning Super Sampling Features

Abstract

With the rise in online shopping, the role of product images in shaping consumer purchase decisions has been accentuated. Despite burgeoning research in this domain, there remains a lacuna in comprehensively understanding the relationship between specific visual attributes, such as color and target shape, in product images and the consequent sales volume. To bridge this gap, the relationship between product image colors and sales volume on online platforms was examined, and color attributes from these images were systematically extracted. Furthermore, an exploration was undertaken into the association between the target shape of product images and sales volume. Deep Learning Super Sampling (DLSS) features from these images were distilled, aiming to furnish a more precise market analysis. Through leveraging advanced machine learning techniques, this study not only augments the academic comprehension of consumer behavior but also proffers strategic insights for online retail practitioners. The methodological approach ensures a targeted marketing direction and facilitates informed product design strategies.