The researchers are working on algorithms that can approximate what an entire image looks like based on only 3 or 4 percent of the information that conventional digital cameras use. This approach would require less compression steps and fewer light detectors, saving power and space. Credit: Digital Signal Processing Group/Rice University