Capability brief
Medical and molecular imaging from raw counts to workflow
Photon-counting CT and nuclear-medicine modalities punish weak detector interfaces and naive readout assumptions. We architect the full thread: sensor and bump/bond strategy, ASIC-friendly pixel maps, timing and coincidence plumbing, and the reconstruction or learning-assisted pipelines that consume the resulting data.
Photon-counting CT and energy-resolved paths
High channel counts, pile-up, and per-pixel calibration drive system-level tradeoffs that cannot be fixed in software after the fact. We work with partners on channelization, rate targets, power envelopes, and data offload—so the front-end matches the gantry or bench system you are actually building.
SPECT, PET, and timing-centric modalities
Nuclear medicine depends on stable timing, coincidence logic, and clean detector coupling. We support detector interfacing, digital timing paths, and integration with reconstruction stacks—whether analytical, iterative, or hybrid ML approaches.
Detector contacts, lifetime, and charge collection
Contact and passivation quality quietly dominates field reliability. We develop and apply strategies—including aluminum-doped zinc oxide (AZO) contact approaches—that aim to extend detector service life and improve charge collection relative to legacy metalizations, especially in demanding radiation environments.
Those material and process choices feed directly into noise budgets, polarization behavior, and long-term gain stability—inputs your reconstruction and AI stages actually see.
Algorithms, simulation, and validation
We implement corrections, iterative reconstruction, and simulation-driven development where ground truth is scarce. GPU and CPU deployment targets are chosen with your throughput and licensing constraints in mind.