Nicholas Howe
Professor of Computer Science — handwriting and document analysis
Northampton, Massachusetts
Summary
Specialist in handwriting and historical document analysis who develops algorithms for document binarization, word spotting, style-based manuscript dating, and related document-image processing tasks. smith+2
Established computer vision and machine learning researcher with earlier work on markerless 3D human motion reconstruction and image retrieval, evidenced by foundational publications and long-standing citation record. smith+1
Experienced undergraduate educator and professor, teaching courses such as web design, data structures, operating systems, and computer vision while maintaining an active research program. smith+1
Engages in interdisciplinary, collaborative projects that apply computer vision and OCR to humanities and regional consortial education efforts, including work on the Krateros squeeze digitization project and Five-College initiatives. ias+1
Work
Education
Projects
Writing
Details of Deformable Part Models for Automatically Georeferencing Historical Map Images
January 1, 2019Presents a probabilistic shape-matching method to align binarized historical map images to GIS vector data, improving georeferencing accuracy for historical map digitization.
Document Binarization with Automatic Parameter Tuning
January 1, 2013Describes a document binarization approach and an automatic technique for tuning key parameters per-image, improving binarization performance on benchmarks and advancing state-of-the-art results.
Bayesian Reconstruction of 3D Human Motion from Single-Camera Video
January 1, 2000Presents a system that reconstructs 3D human motion from single-camera video by using learned priors about human motion to resolve 3D ambiguities; demonstrates automatic tracking and reconstruction on multiple sequences.