Using R to determine the impact of an image on viewers

Nicolae-Marius JULA (
University of Bucharest, Faculty of Business and Administration
Diana Mihaela JULA, PhD student (
The Romanian Academy


In this paper, we are addressing some aspects regarding the likeability of an image. Besides the obvious answers as the quality, resolution, size, we suggest that there are also other properties that make people like and remember an image. This study answers the questions like: what is the best time frame to post on social media for the highest impact, is the dominant color, main subject, or the number of subjects having an impact on clients/followers?
We are answering these questions using AI packages like YOLO for classification and object detection, countcolors for findind pixels by color range, and imager for an array of functions for working with image data.

Keywords: image processing, color detection, object detection

[Full Text]

Romanian Statistical Review 2/2022