Our Case Studies
Explore our successful data science projects and solutions
Historical Newspaper Article Deduplication
Using image embeddings to identify duplicate articles across digitized newspaper archives.
OCR Extraction Runtime Optimization
Improved AWS Lambda OCR processing speed by 100x using computational geometry techniques.
Accurate Line Detection for Historical Documents
Advanced line segmentation for improved OCR accuracy on degraded historical texts.
Newspaper Column Detection via Fourier Analysis
Automatically detecting newspaper columns using spectral analysis to improve OCR layout understanding.
Large-Scale Face Clustering for Photo Archives
Grouping millions of historical portrait photos by identity using deep learning embeddings.
Automated Restoration of Degraded Historical Images
Restoring faded, scratched, and stained historical photographs using deep learning-based inpainting and colorization.
Robust Page Detection for Scanned Historical Books
Automatically extracting page boundaries from warped, unevenly lit historical book scans.
Precise Newspaper Article Segmentation for Digital Archives
Detecting and isolating individual articles in dense, multi-column historical newspaper layouts.