Unlimited sensing is a novel digital sensing paradigm that tries to overcome the inherent trade-offs and limitations of conventional analog-to-digital converters (ADC). See this SIAM News Blog for a comprehensive introduction of the main ideas.
Project description
Unlimited sensing promises to be particularly advantageous in applications where some components of the signal have a large amplitude while others, which are also important, have a relatively small amplitude and will be sensed with high digitization noise using conventional ADCs.
Photoacoustic tomography in biological tissues (e.g., for breast imaging) is such an application: due to the exponential fluence decay in the tissue, signals originating from deeper anatomical structures will be significantly weaker than those originating from the skin layer and in the reconstructed image, digitization noise will impact those regions more heavily. In this project, we will first demonstrate these effects in a simulation study. Then, we will examine if using an unlimited sensing framework would be able to reduce the digitization noise without introducing other sampling errors and improve the reconstructed image.
Supervision & focus areas
Supervisor: Felix Lucka
Keywords: unlimited sensing, signal processing, photoacoustic tomography, inverse problems
