Fangda Li

Fangda Li, Ankit Manerikar, Tanmay Prakash et al.

When dealing with material classification in baggage at airports, Dual-Energy Computed Tomography (DECT) allows characterization of any given material with coefficients based on two attenuative effects: Compton scattering and photoelectric absorption. However, straightforward projection-domain decomposition methods for this characterization often yield poor reconstructions due to the high dynamic range of material properties encountered in an actual luggage scan. Hence, for better reconstruction quality under a timing constraint, we propose a splitting-based, GPU-accelerated, statistical DECT reconstruction algorithm. Compared to prior art, our main contribution lies in the significant acceleration made possible by separating reconstruction and decomposition within an ADMM framework. Experimental results, on both synthetic and real-world baggage phantoms, demonstrate a significant reduction in time required for convergence.

Fangda Li, Ankit V. Manerikar, Avinash C. Kak

Motivated by what is required for real-time path planning, the paper starts out by presenting sRMPD, a new recursive "local" planner founded on the key notion that, unless made necessary by an obstacle, there must be no deviation from the shortest path between any two points, which would normally be a straight line path in the configuration space. Subsequently, we increase the power of sRMPD by using it as a "connect" subroutine call in a higher-level sampling-based algorithm mRMPD that is inspired by multi-RRT. As a consequence, mRMPD spawns a larger number of space exploring trees in regions of the configuration space that are characterized by a higher density of obstacles. The overall effect is a hybrid tree growing strategy with a trade-off between random exploration as made possible by multi-RRT based logic and immediate exploitation of opportunities to connect two states as made possible by sRMPD. The mRMPD planner can be biased with regard to this trade-off for solving different kinds of planning problems efficiently. Based on the test cases we have run, our experiments show that mRMPD can reduce planning time by up to 80% compared to basic RRT.