This repository hosts instrument control software for the Remoscope, an automated and low-cost malaria diagnostic system, requiring minimal user training. The Remoscope was developed by the Bioengineering team at the Chan Zuckerberg Biohub San Francisco (CZBSF). The Remoscope images unstained fresh whole blood in liquid form and can screen up to two million red blood cells for Plasmodium falciparum parasites.
Please see our preprint here:
The Remoscope runs off a Raspberry Pi 4 Model B 8GB and an Intel Neural Compute Stick 2 (no longer available) to accelerate on-device neural network inferencing. The instrument control software is written in Python 3.7.3.
├── ulc_mm_package
│ ├── utilities
│ ├── summary_report
│ ├── neural_nets
│ ├── image_processing
│ ├── hardware
│ ├── configs
│ ├── QtGUI
│ ├── scope_constants.py
├── OS_instructions
│ ├── OS_setup.md
│ └── OS_flash.md
├── architecture_description
├── setup.py
└── README.md
For a more detailed description of each folder and its scripts, see the architecture_description/ folder.
NOTE: If you are using a version downstream of commit 283a823 on develop (2024-03-06), you will need to install the stats_utils module. To do so, in the root directory run
python3 -m pip install -e .
- Flash a pre-existing image to an SD card (must be at least 32GB), see
OS_flash.md
- Follow the instructions under
OS_instructions.
To start the software, navigate to ulc_mm_package/QtGUI and run python3 oracle.py.
To use developer mode, navigate to ulc_mm_package/QtGUI and run python3 dev_run.py. This opens a GUI with manual hardware control for debugging purposes.
- Object detection - YOGO: https://github.com/czbiohub-sf/yogo
- Autofocus - SSAF: https://github.com/czbiohub-sf/ulc-malaria-autofocus
- Data processing utilities / model evaluations: https://github.com/czbiohub-sf/lfm-data-utilities/
- Statistics utilities: https://github.com/czbiohub-sf/remo-stats-utils