Imaging is indispensable for nearly every field of science, engineering, technology, and medicine. However, measurement noise and stochastic distortions pose fundamental limits to accessible spatiotemporal information despite impressive tools such as SIM, PALM/STORM, and STED microscopy. How to combat this challenge ideally has been an open question for decades. Inspired by a "virtual gain" technique to compensate losses in metamaterials, "active convolved illumination" has been recently proposed to significantly improve the signal-to-noise ratio, hence data acquisition. In this technique, the light pattern of the object is superimposed with a correlated auxiliary pattern, the function of which is to reverse the adverse effect of noise and random distortion based on their spectral characteristics. Despite enormous implications in statistics, an experimental realization of this novel technique has been lacking to date. Here, we present the first experimental demonstration. We find that the active convolved illumination does not only boost the resolution limit and image contrast, but also the resistance to pixel saturation. The results confirm the previous theories and opens up new horizons in a wide range of disciplines from atmospheric sciences, seismology, biology, statistical learning, and information processing to quantum noise beyond the fundamental boundaries.