Skip to content

Latest commit

 

History

History
68 lines (40 loc) · 3.94 KB

research.md

File metadata and controls

68 lines (40 loc) · 3.94 KB
layout title permalink order
page
Research
/research/
2

PhD Thesis

Making large art historical photo archives searchable

B. Seguin, PhD Thesis, 2018

![Thumbnail]({{ site.baseurl }}/uploads/publications/thesis_thumbnail.png){: .align-center width="80%"}

Selected publications related to my PhD Project

The Replica Project: Building a visual search engine for art historians

B. Seguin, XRDS: Crossroads, The ACM Magazine for Students - Computers and Art, 2018

Invited journal article describing the high level ideas behind the Replica project.

![Thumbnail]({{ site.baseurl }}/uploads/publications/XRDS_thumbnail.png){: .align-center width="60%"}

Visual Link Retrieval in a Database of Paintings

B. Seguin, C. Striolo, I. diLenardo, F. Kaplan, ECCV Visart Workshop 2016

Technical paper which forms the basis of the learning of the visual similarities for the Replica search engine.

![Thumbnail]({{ site.baseurl }}/uploads/publications/Visart_thumbnail.png){: .align-center width="70%"}

[New Techniques for the Digitization of Art Historical Photographic Archives—the Case of the Cini Foundation in Venice]({{ site.baseurl }}/uploads/publications/Archiving2018.pdf)

B. Seguin, L. Costiner, I. diLenardo, F. Kaplan, Archiving 2018

Describes the processing pipeline used for the digitization and automatic processing of the Cini photo-collection.

![Thumbnail]({{ site.baseurl }}/uploads/publications/Archiving_thumbnail.png){: .align-center width="70%"}

dhSegment: A generic deep-learning approach for document segmentation

B. Seguin*, S. Oliveira*, F. Kaplan, Frontiers in Handwriting Recognition (ICFHR) 2018 (website+code)

How the deep-learning approach we used originally for the processing of the Cini collection was generalized to many other cases of document processing for the Venice Time Machine.

![Thumbnail]({{ site.baseurl }}/uploads/publications/dhSegment_thumbnail.png){: .align-center width="60%"}

A Learning Interface for Finding Visual Connections in Artworks

B. Seguin, L. Costiner, I. diLenardo, F. Kaplan, 2017 (to be released soon)

Finding visual connections between artworks is difficult because it relies on the exploration of very large corpuses of images. Here we present how one can continuously leverage previously acquired connections in order to help users explore the image space more efficiently.

![Thumbnail]({{ site.baseurl }}/uploads/publications/interface_thumbnail.png){: .align-center width="60%"}

Other work

This is an unrelated work with my main project that came out of a fun and intense collaboration with two friends on applying deep-reinforcement-learning to logic optimization.

Deep Learning for Logic Optimization Algorithms

W. Haaswijk*, E. Collins*, B. Seguin*, M. Soeken, S. Süsstrunk, F. Kaplan, S. De Micheli, International Symposium on Circuits and Systems 2018.

![Thumbnail]({{ site.baseurl }}/uploads/publications/logic_thumbnail.png){: .align-center width="60%"}