HEAP

HEAP automates identification and quantification of microscopic helminth eggs using deep learning to support parasitology diagnostics.

Key Features:

• Model architectures: Implements SSD (Single Shot MultiBox Detector), U-net, and Faster R-CNN (Faster Region-based Convolutional Neural Network).
• Model selection: Allows selection of the most suitable prediction model for a given specimen or analysis task.
• Image binning and egg-in-edge algorithm: Employs pixel density detection to enhance microscopic image resolution and improve egg identification performance.
• Distributed computing: Supports distributed computing across various operating systems to reduce computation time and operate on low-cost computers.
• Parasitic egg database: Includes a collection of microscopic helminth egg images with associated labeling data and pretrained models.
• Automated quantification: Provides automated counting and quantification of helminth eggs from microscopic specimens.

Scientific Applications:

• Parasitology diagnostics: Automated image-based detection and counting of helminth eggs to support diagnostic workflows.
• Research and model development: Uses labeled image datasets and pretrained models for development and validation of deep learning approaches in parasitology.
• Training and education: Serves as a resource for training technicians and students in microscopic helminth egg examination.
• Field and low-resource screening: Enables deployment of automated egg detection and quantification on low-cost computers in resource-limited laboratory settings.

Methodology:

Integrates SSD, U-net, and Faster R-CNN architectures; applies image binning and an egg-in-edge algorithm using pixel density detection; operates via distributed computing across multiple operating systems and provides labeled datasets and pretrained models.