Movement Artifact Direction Estimation Based on Signal Processing Analysis of Single-Frame Images.

Movement artifact direction and magnitude are critical parameters in noise detection and image analysis, especially for single-frame images where temporal information is unavailable. This paper introduces the Movement Artifact Direction Estimation (MADE) algorithm, a signal processing-based approach that performs 3D geometric analysis to estimate both the direction (in degrees) and weighted quantity (in pixels) of movement artifacts. Motivated by computational challenges in medical image quality assessment systems such as LUIAS, this work investigates directional multiplicative noise characterization using controlled experimental conditions with optical camera imaging. The MADE algorithm operates on multi-directional quantification outputs from a preprocessing pipeline-MAPE, ROPE, and MAQ. The methodology is designed for computational efficiency and instantaneous processing, providing interpretable outputs. Experimental results using precision-controlled apparatus demonstrate robust estimation of movement artifact direction and magnitude across a range of image shapes and velocities, with principal outputs aligning closely to ground truth parameters. The proposed MADE algorithm offers a methodological proof of concept for movement artifact analysis in single-frame images, emphasizing both directional accuracy and quantitative assessment under controlled imaging conditions.