Pixon-based multiresolution image reconstruction for Yohkoh's hard X-ray telescope

We present results from the application of pixon-based multiresolution image reconstruction to real and simulated data from Yohkoh's Hard X-ray Telescope (HXT). The goal of the pixon algorithm is to minimize the number of degrees of freedom used to describe an image within the accuracy allowed by the noise. This leads to a reconstruction that is optimally constrained. We apply the pixon code to two solar flares in the HXT database and compare the results of the pixon reconstruction to the results of a direct, linear, smoothed inversion of the HXT Fourier synthesis data and to a maximum entropy reconstruction. The maximum entropy reconstruction is vastly better than the direct inversion, but the pixon reconstruction gives superior noise suppression and photometry. Further, the pixon reconstruction does not suffer from overresolution of the images.