Techniques for estimating and predicting CPU resource availability during task execution in single- and multi-core systems are presented. This paper introduces a new task execution and performance prediction model to enhance the efficiency of image steganography overhead. The main goal of steganography is to make embedded secret data undetectable to both computational and human analysis. This paper provides a brief overview of 2D and 3D image steganography. A new, concise survey is necessary as many novel techniques have been developed since the last review, and previous surveys did not cover 3D image steganography. It also briefly discusses the spatial, JPEG, and side-informed JPEG domains in 2D image steganography, along with the geometrical, topological, and representation spatial domain techniques used in 3D steganography. The analytical models proposed here are useful for predicting CPU resource availability during the execution of new image steganography tasks. Extensive experimental studies with real-world benchmark programs and statistical analysis are conducted to evaluate the accuracy of the suggested empirical models.