Skip to main content

Automatic Speech Analysis for Cognitive Assessment

Project description

ASACA – Automatic Speech Analysis for Cognitive Assessments

CI PyPI License Python GUI

ASACA is an end-to-end toolkit that transforms raw speech into multimodal biomarkers — lexical, prosodic and pause-based — and returns an interpretable prediction ( HC / MCI / AD ) and low Word error rate transcriptions (WER <0.02)).


✨ Key Features

Capability Detail
Single-command inference asaca run audio.wav outputs JSON + PDF report
Fine-tuned wav2vec 2.0 ASR < 2 % WER on in-domain test set
Explainability SHAP plots per classification
Rich feature set word-error rate, syllable rate, pause stats, spectral cues
Offline-ready Model weights stored under Models/ via Git LFS
PEP 517/621 packaging pip install asaca or editable mode

🚀 Quick start

Install the package from PyPI and run inference on a WAV file:

pip install asaca
asaca-cli infer path/to/audio.wav -o out/

Alternatively install in editable mode for development:

git clone https://github.com/RhysonYang-2030/ASACA-Automatic-Speech-Analysis-for-Cognitive-Assessment.git
cd ASACA-Automatic-Speech-Analysis-for-Cognitive-Assessment
pip install -e .[dev]

The CLI outputs recognised text along with a PDF report and JSON file in the specified output directory.

Usage

Pipeline

asaca/
├── src/             # library code
├── tests/           # unit tests
├── docs/            # MkDocs documentation
├── examples/        # example notebooks and data
└── notebooks/       # tutorial notebooks

Run asaca-cli --help to see all commands including feature extraction.

Documentation

Full API reference and user guide live in the docs/ directory and on Read the Docs.

Docker

Build the image and run inference in an isolated environment:

docker build -t asaca .
docker run --rm -v "$PWD:/data" asaca asaca-cli infer /data/audio.wav

The container entrypoint defaults to asaca-cli.

License

Released under the Apache-2.0 license.

Citation

If you use ASACA in your research, please cite the project using the CITATION.cff file.

Contact

Maintainer: Xinbo Yang

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

asaca-0.1.0.tar.gz (47.4 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

asaca-0.1.0-py3-none-any.whl (53.9 kB view details)

Uploaded Python 3

File details

Details for the file asaca-0.1.0.tar.gz.

File metadata

  • Download URL: asaca-0.1.0.tar.gz
  • Upload date:
  • Size: 47.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.12.9

File hashes

Hashes for asaca-0.1.0.tar.gz
Algorithm Hash digest
SHA256 e6bf7ad747fd70f2824ab8eb9437064471b99bed2055915cffa2d150591791fb
MD5 a26a719f8131a27c9545231d616e25c1
BLAKE2b-256 46c75ebdb42c7ef7438b5e21e02a02fe9af26a48b35fcf5629a6e380a2b991d1

See more details on using hashes here.

File details

Details for the file asaca-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: asaca-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 53.9 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.12.9

File hashes

Hashes for asaca-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 1dc9fe6db3df059b2132b80ceb0c5c966223bde6002f876b174957e376df6793
MD5 a0c81a56cc091a42729d4b51d391e37b
BLAKE2b-256 b0ab1fa0e74b1341c276d39dfac56ef54ab009f508c0e115ba1d5f3f1b43e5c1

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page