This paper presents a GPU-based method for denoising of time series images obtained by confocal microscopy, in order to study the oscillations of calcium and membrane voltage potential in beta cells. Since denoising of captured images is one of the first crucial steps, it is desirable that it is highly efficient and fast. This is especially important when dealing with a large number of time series images, where the computational complexity increases tremendously. Results demonstrate that the proposed method is at least 5 times faster, in comparison with a CPU-based implementation, while retaining the same accuracy.