As conventional hardware and algorithms are being driven to their limits, researchers have started to realize the benefits of a co-design approach that leverages on a symbiotic relationship between hardware and algorithms. Instead of traditional “capture first, process later” approach, the hardware-software co-design approach surpasses traditional bottlenecks in sensing. At the convergence of inverse problems, physical models and optics/vision, the module will focus on the foundational principles of this emerging topic. Theoretical approaches will focus on abstracting inverse problems from wider, inter-disciplinary applications and provide for mathematical tools for solving such problems. Practical/applied experience will entail application of algorithms on experimental data acquired from novel sensors.

By the end of this module, you will be able to: - model the data acquisition process by using advanced mathematical tools - develop algorithms for inverse problems. - design alternative sensors by leveraging tools from different domains. - analytically, numerically and experimentally analyse data from novel sensor.

Introduction to Computational Sensing Sensing and Imaging Toolkit: a) 1D Sampling | 2D Imaging | 3D Voxel Acquisition b) Optics, Image Sensors and Image Formation Computational Toolkit: a) Modeling: Forward vs Inverse Problems b) Sampling Architectures c) Math Tools (Signal Proc., Linear Algebra and Optimization). d) Algorithmic Toolkit (e.g. Spectral Estimation, Orthogonal Matching Pursuit, Super-resolution, One-Bit Noise Shaping, Sparse Recovery etc.) Plenoptic Coding: a) Spatially Coded Imaging b) Temporally Coded Imaging c) Light-field Imaging d) Multispectral Imaging Case Studies a) Single-Pixel Imaging b) Time-of-Flight Imaging c) Fluorescence Lifetime Imaging d) Modulo Sensing and Imaging e) Tomography f) One-Bit Sensing