John Hyunchul Hong

Continuous color reflective displays using interferometric absorption

Reflective displays that rely on ambient light as opposed to an internal light source have been making inroads for a variety of important applications, especially those involving mobility where power usage must be aggressively controlled. The underlying color rendering strategies for both reflective and emissive displays have largely been the same, combining three fixed, primary color subpixels to compose the rich gamut that users expect. The result, for reflective color displays, is unfavorable brightness/gamut performance since each of the color subpixels absorbs roughly 2/3 of the incident white light. We demonstrate a new technology that we call the single mirror interferometric display that overcomes such limitations with pixels whose reflectance properties can tune through a continuum of colors, including high contrast black and white reflectance states. We use an effect that we call interferometric absorption in which a thin absorbing metal layer in front of a highly reflective mirror surface selectively absorbs different colors, depending on the gap that separates the two. The gap is controlled by electrostatic actuation in a relatively simple micro-electro-mechanical-system structure. We describe this elegant and powerful color rendering principle and present experimental results for the basic pixel device as well as early system demonstrations.

Extension of FRI for modeling of electrocardiogram signals

Recent work has developed a modeling method applicable to certain types of signals having a “finite rate of innovation” (FRI). Such signals contain a sparse collection of time- or frequency-limited pulses having a restricted set of allowable pulse shapes. A limitation of past work on FRI is that all of the pulses must have the same shape. Many real signals, including electrocardiograms, consist of pulses with varying widths and asymmetry, and therefore are not well fit by the past FRI methods. We present an extension of FRI allowing pulses having variable pulse width (VPW) and asymmetry. We show example results for electrocardiograms and discuss the possibility of application to signal compression and diagnostics.

Keypoint clustering for robust image matching

A number of popular image matching algorithms such as Scale Invariant Feature Transform (SIFT)1 are based on local image features. They first detect interest points (or keypoints) across an image and then compute descriptors based on patches around them. In this paper, we observe that in textured or feature-rich images, keypoints typically appear in clusters following patterns in the underlying structure. We show that such clustering phenomenon can be used to: 1) enhance recall and precision performance of the descriptor matching process, and 2) improve convergence rate of the RANSAC algorithm used in the geometric verification stage.