Heat release rate (HRR) is a fundamental quantity in combustion, as it characterizes the spatial and temporal distribution of energy release within a flame. It plays a central role in flame dynamics and is a primary driver of thermo-acoustic coupling and combustion instabilities, which can have a detrimental effect on combustion systems[1, 2]. Despite its importance, heat-release rate cannot be directly measured experimentally. Currents methods rely on band-pass filtered measurements of OH∗ or CH∗ radicals formed in key exothermic reaction steps. However the signal is inherently indirect as it is a superposition of desired radical emissions and broadband contributions from other species, and is highly sensitive to local equivalence ratio and turbulence intensity in the reaction zone. To address this limitation, we develop a latent diffusion model[3] capable of prediction spatially-resolved HRR fields from experimental chemiluminescence and schlieren images. Unlike deterministic regression models, the probabilistic nature of diffusion models naturally accounts for the inherent stochasticity of turbulent flames, enabling the model to capture the range of possible HRR realizations consistent with given experimental observations. This is done under the assumptions that high fidelity Large Eddy Simulations (LES) are good approximations to reality hence, models can be trained on simulations and be deployed experimentally. We first validate the approach on a 2D. advection-diffusion system with known source, then extended it to reacting LES. Preliminary results will be shown on experimental data demonstrating the model’s potential for enabling quantitative HRR measurements in practical combustion systems. [1] Phillip H. Paul and Habib N. Najm. Planar laser-induced fluorescence imaging of flame heat release rate. Symposium (International) on Combustion, 27(1):43–50, January 1998. [2] B.O. Ayoola, R. Balachandran, J.H. Frank, E. Mastorakos, and C.F. Kaminski. Spatially resolved heat release rate measurements in turbulent premixed flames, Combustion and Flame, 144(12):1–16, January 2006. [3] Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj¨orn Ommer. High Resolution Image Synthesis with Latent Diffusion Models, April 2022. arXiv:2112.10752