Photograph Quality Prediction using Convolutional Neural Networks

Ryoma Okuno, Yasuaki Ito, Koji Nakano


In this paper, we propose a machine learning method to evaluate the quality of photographs with high accuracy. Specifically, the assessment of photographs is performed as a problem of classifying whether the input photograph was taken by an amateur or an expert photographer. The proposed network model utilizes ConvNext, a convolutional neural network, and provides histograms as auxiliary input as color distribution information to improve the accuracy. The experimental results show that the proposed model achieves a test accuracy of 92.5%. 


photograph quality prediction; CNN; ConvNext; machine learning

Full Text:



  • There are currently no refbacks.