Pattern spotting in historical documents using convolutional models
Professor Advisor
dc.contributor.advisor
Saavedra Rondo, José
Professor Advisor
dc.contributor.advisor
Ríos Pérez, Sebastián
Author
dc.contributor.author
Úbeda Soto, Ignacio Andrés
Associate professor
dc.contributor.other
Heutte, Laurent
Associate professor
dc.contributor.other
Sauré Valenzuela, Denis
Admission date
dc.date.accessioned
2020-10-15T02:27:39Z
Available date
dc.date.available
2020-10-15T02:27:39Z
Publication date
dc.date.issued
2020
Identifier
dc.identifier.uri
https://repositorio.uchile.cl/handle/2250/177136
General note
dc.description
Tesis para optar al grado de Magíster en Gestión de Operaciones
es_ES
General note
dc.description
Memoria para optar al título de Ingeniero Civil Industrial
Abstract
dc.description.abstract
Pattern spotting consists of locating different instances of a given object (i.e. an image query) in a collection of document images. Contrary to object detection, no prior information is given about the patterns that can be searched, hence no training can be done for localization. The queried patterns may vary in shape, size, color, context and even style, which makes pattern spotting a difficult task. To tackle this problem, we propose a convolutional neural network approach based on Feature Pyramid Networks (FPN) as the feature extractor of our system. Using FPN allows us to extract descriptors of local regions of the documents to be indexed and queries, at multiple scales with a single forward pass. Our main hypothesis is that deep features are more discriminatory than classical descriptors for pattern localization. Experiments conducted on DocExplore (a historical document dataset for pattern spotting evaluation) show
that the proposed system improves mAP by 73% (from 0.157 to 0.272) in pattern localization compared with state-of-the-art results, even when the feature extractor is not trained with domain-specific data. Memory requirement and computation time are also decreased since the descriptor dimension used for distance computation is reduced by a factor of 16. We conclude that CNN-based local descriptors are better than VLAD (classical) descriptors at locating patterns and we use them to propose a system for pattern localization. Limitation of our approach is that it struggles with non-square patterns. We also propose a solution to address this issue extracting multiple descriptors per query. Although it improves results in retrieving documents, it loses precision in locating patterns. Aggregation on those descriptors is proposed as interesting future
work in order to improve the system.