FRET has emerged as an important tool for studying intracellular processes and interactions between biomolecules. Intracellular donor and acceptor molecules are distributed in individual organelles that usually have complex non-continuous shape. Consequently, background pixels arising from fluorophore-free regions of the cell are proximal to FRET-positive pixels, leading to systemic errors in the estimated FRET values. This study introduces a new FRETTH algorithm for FRET estimation by acceptor photobleaching that separates the FRET-positive pixels from the background by applying user-defined thresholds for pixel selection. The FRETTH algorithm was validated by analysis of interactions between fluorescently tagged proteins in the endoplasmic reticulum using acquired and simulated images. The novel algorithm showed superior performance to the regular FRET calculation algorithm in acquired images and in most simulations. The developed algorithm was incorporated into the FRETcalc plugin for ImageJ program that enables user-defined choices of thresholds for calculation of FRET by acceptor photobleaching. (c) 2007 Elsevier Inc. All rights reserved.