Automated vector selection of SIVQ and parallel computing integration MATLAB™: Innovations supporting large-scale and...

Area of Science:

• Digital pathology
• Computational imaging
• Biomedical image analysis

Background:

• Spatially invariant vector quantization (SIVQ) is a texture and color-based image matching algorithm.
• Previous SIVQ applications required manual vector selection, a time-consuming process.
• SIVQ's inherent scalability suggested potential for high-throughput computing.

Purpose of the Study:

• To develop an automated process for selecting optimal ring vectors in SIVQ.
• To enhance the efficiency and accessibility of SIVQ for histopathological feature identification.
• To explore SIVQ's high-throughput capabilities.

Main Methods:

• Developed an automated ring vector selection process using receiver operating characteristic (ROC) analysis.
• Candidate vectors were generated and compared against user-defined ground truth regions.
• Area-under-the-curve (AUC) was used as a metric for vector goodness-of-fit.

Main Results:

• Successfully identified optimized vectors for identifying malignant colonic epithelium and soft tissue sarcoma.
• The automated selection process yielded satisfactory results as defined by the AUC metric.
• Integrated SIVQ with MATLAB™ for enhanced computational throughput.

Conclusions:

• The SIVQ algorithm is effective for automated vector selection in histopathology.
• SIVQ demonstrates suitability for high-throughput computational settings.
• Automated SIVQ offers a scalable solution for image analysis in digital pathology.