fairfolio
Saved Projects

VirtualPaL: Novel IDC Segmentation & Classification in Whole Slide Imaging

AJAS · 2026

Harmon Hong · Texas Academy of Science

Overview

Breast cancer is the most common cancer type among women with IDC (Invasive Ductal Carcinoma), accounting for 80% of cases. Living in the age of precision medicine, WSI (Whole Slide Imaging) offers high resolution, tissue-based microscopic digital pathology images, producing greater accuracy and efficiency compared to other commonly used diagnostic procedures. With the onset of AI in a medical context, VirtualPaL utilizes deep learning to assist clinical professionals, cancer researchers, and affected patients via IDC-specific WSI visualization, analysis, segmentation, and diagnostic support. The dataset used to train VirtualPaL’s breast cancer image analysis model was curated by researchers at Case Western Reserve University and later uploaded to the public library, Kaggle. The data consisted of 162 whole mount slide images of potential Breast Cancer patients, from which 50x50 image patches were extracted, creating 198,738 IDC negative and 78,786 IDC positive patches. VirtualPaL runs a novel model for classification, which was developed with the deep learning CNN model, ResNet-50, as its base, with fine-tuned custom layers implemented to optimize classification. When processing images, VirtualPaL uses an innovative dynamic patch extraction logic for enhanced pattern recognition and focus on small features or anomalies within the fine tissue. The first of its kind, VirtualPaL specializes in WSI analysis of IDC and is available to the public. After optimization, VirtualPaL’s custom model reached a final validation accuracy of 95.1%. Precision, recall, and AUC scores were 91.5%, 91%, and 98%, respectively, showing a favorable reduction of false negatives and false positives in comparison to current studies in the field and achieved the predefined performance target. With its accuracy and metrics, the model is capable of predictive diagnosis with human approval. Along with VirtualPaL’s image processing and visualization ability, VirtualPaL presents an effective solution to improve the accuracy and efficiency of IDC pathology analysis.

Competition history

  • AJAS 2026 Category not listed

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

Source: AAAS Annual Meeting (Confex) / American Junior Academy of Science

Save projects to your library

Sign in with Google to keep track of projects you find interesting, organized into folders. An account also raises your daily allowance for “Has this been done?”, and lets you create a key for the MCP server with a much higher limit than anonymous use. Browsing stays public.

Continue with Google