---
id: aot-biomaps
version: "2.9.905"
license: unclear
license_treatment: unclear
maintenance: active
---

# aot-biomaps — Acousto-Optic Tomography Reconstruction Library

License: unclear · Maintenance: active · Downloads: 124.3K/mo

## What it is and what it does

AOT_biomaps is a specialized library for Acousto-Optic Tomography, a biomedical imaging modality that combines acoustic and optical physics. It implements multiple tomographic reconstruction algorithms (MLEM, PDHG, LS, DEPIERRO, MAPEM, LBFGS) with both CPU (NumPy) and GPU (CuPy) backends, allowing automatic fallback between implementations. The library also provides acoustic wave simulation, optical media modeling, signal processing utilities, and 2D/3D visualization.

The package is designed for research in biomedical imaging and combines advanced reconstruction mathematics with practical optimization for sparse matrix operations. It depends on numpy, scipy, tqdm, matplotlib, and optionally cupy-cuda12x for GPU acceleration and k-wave-python for acoustic simulation. Installation is straightforward via pip, with a CPU-only mode available for systems without CUDA.

Use it for:

- Reconstruct 3D biomedical images from acousto-optic tomography sensor data using iterative algorithms like MLEM or PDHG.
- Simulate acoustic wave propagation and optical interactions in heterogeneous tissue models for AOT experiment design.
- Accelerate tomographic reconstruction on GPU hardware to reduce computation time from minutes to seconds.
- Process and visualize 2D/3D tomographic datasets with built-in filtering, backprojection, and Radon transform tools.
- Benchmark different sparse matrix formats (CSR, SELL-C-sigma) for memory efficiency in large-scale reconstruction problems.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

AOT_biomaps is a Python library for Acousto-Optic Tomography image reconstruction, providing tomographic algorithms (MLEM, PDHG, LS, DEPIERRO, MAPEM, LBFGS) with CPU and GPU implementations, plus acoustic simulation, optical modeling, and signal processing.

Yes, if you are working on acousto-optic tomography research or biomedical imaging reconstruction. The library is actively maintained, has low install friction, offers both CPU and GPU paths, and implements multiple state-of-the-art algorithms. However, verify the license status before use in proprietary work, and confirm that k-wave-python is available if acoustic simulation is required. Not relevant for general-purpose image processing or non-AOT applications.

## Install

pip install aot-biomaps
uv add aot-biomaps
poetry add aot-biomaps

## Installing aot-biomaps

Before you install: Low install friction; pure Python wheel with straightforward pip installation. Active maintenance (latest release 16 days ago). Six runtime dependencies including numpy, scipy, and optional GPU support via cupy-cuda12x; CPU-only installation available via environment variable.

License in practice: License treatment is unclear—the description mentions MIT but the fact sheet contains no SPDX identifier or confirmed license metadata. Verify the actual license before use in proprietary or restricted contexts.

Quickstart:

pip install aot-biomaps

import numpy as np
from AOT_biomaps import Tomography, AlgebraicRecon
from AOT_biomaps.AOT_Recon.ReconEnums import ReconType

experiment = Tomography(
    optic_image_path="path/to/optic_image.npy",
    acoustic_fields_path="path/to/acoustic_fields.npy"
)
recon = AlgebraicRecon(
    experiment=experiment,
    reconType=ReconType.Algebraic,
    optimizerType="MLEM",
    numIterations=100
)
recon.run(withTumor=True)

Requires Python >=3.8. GPU support (cupy-cuda12x) requires CUDA 12.x; CPU-only mode available via AOT_BIOMAPS_CPU_ONLY=true environment variable or --cpu flag.

Verify before relying:

- Whether MIT license in description is the actual license or if license metadata will be clarified in PyPI
- Whether k-wave-python is truly optional or required for acoustic simulation features
- Actual performance benchmarks (21.5x GPU speedup claimed) on typical user hardware
- Whether sparse matrix implementations (CSR, SELL-C-sigma) are production-ready or experimental

## Package facts

- License: not declared (unclear)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 124.3K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

acousto-optic tomography reconstruction, tomographic image reconstruction algorithms, GPU accelerated tomography, MLEM PDHG reconstruction, biomedical imaging library, acoustic simulation python, sparse matrix tomography, biomedical-imaging, gpu-accelerated, tomography

[View on SkillFed](https://skillfed.io/packages/aot-biomaps) · [View on PyPI](https://pypi.org/project/aot-biomaps/)