Maxwell McInnisStart a project

Case study

SPECT Image Reconstruction

A Python-based implementation of SPECT image reconstruction for nuclear medicine imaging. The project loads emission and attenuation CSV profiles, computes attenuation correction factors, reconstructs cross-sectional slices using filtered back projection, and performs statistical validation of the output — mirroring the computational pipeline used in clinical SPECT/CT systems.

RoleMedical image reconstruction, scientific computing, data analysis
TimelineUniversity project — ENGG4040
FocusMedical Imaging + Scientific Computing
Image reconstructionAttenuation correctionNuclear medicine imaging
Nuclear medicine / SPECT/CTDomain
FBP · attenuation correction · sinogramMethods
https://github.com/MaxwellM34/ENGG4040_Lab4_SPECT_RECONSTRUCTIONSource

The brief

Challenge

Reconstruct a meaningful cross-sectional image from raw 1D emission projection profiles, accounting for photon attenuation through tissue.

Approach

What we made

Loaded emission and attenuation CSV profiles, computed per-ray attenuation correction factors, then applied filtered back projection to synthesize reconstructed slices. Output was validated statistically against known phantom values.

  • Implemented attenuation correction factor (ACF) computation from raw emission and attenuation profiles.
  • Reconstructed SPECT cross-sectional slices using filtered back projection methodology.
  • Validated outputs with statistical analysis (mean and standard deviation) against known phantom geometry.

Outcome

Results

Successfully reconstructed SPECT slices with attenuation correction, producing results consistent with expected phantom geometry and demonstrating the core computational pipeline of clinical nuclear medicine systems.

PythonNumPyMatplotlibCSV

Gallery

Visual snapshots

Click any image to expand.

Next project

Autonomous Maze Vehicle

An Arduino-powered autonomous vehicle that navigates mazes using three ultrasonic sensors, proportional wall-following control, and real-time obstacle avoidance — all on a resource-constrained microcontroller.

View next project