Publications
2018
Saikia, Surajit; Fidalgo, Eduardo; Alegre, Enrique; Fernández-Robles, Laura
Query based object retrieval using neural codes Artículo de revista
En: International Joint Conference SOCO’17-CISIS’17-ICEUTE’17 León, Spain, September 6–8, 2017, Proceeding 12, pp. 513–523, 2018, (Publisher: Springer International Publishing).
Resumen | Enlaces | BibTeX | Etiquetas: Cosine Similarity, Faster R-CNN, Neural Codes, Object Retrieval
@article{saikia_query_2018,
title = {Query based object retrieval using neural codes},
author = {Surajit Saikia and Eduardo Fidalgo and Enrique Alegre and Laura Fernández-Robles},
url = {https://link.springer.com/chapter/10.1007/978-3-319-67180-2_50},
year = {2018},
date = {2018-01-01},
journal = {International Joint Conference SOCO’17-CISIS’17-ICEUTE’17 León, Spain, September 6–8, 2017, Proceeding 12},
pages = {513–523},
abstract = {This paper proposes an object retrieval method that leverages the neural codes (activations) generated by the last inner-product layer of the Faster R-CNN network, showing its effectiveness not just for object detection but also for retrieval tasks. The method processes a subset of ImageNet images, storing neural codes, object information, and bounding box data for fast retrieval. Given a query image, the system identifies and retrieves objects by calculating cosine similarity against saved neural codes, achieving efficient retrieval with a time of 0.534 seconds for 1454 objects in the test set.},
note = {Publisher: Springer International Publishing},
keywords = {Cosine Similarity, Faster R-CNN, Neural Codes, Object Retrieval},
pubstate = {published},
tppubtype = {article}
}
2017
Fernández-Robles, Laura
Recognition and retrieval of objects in diverse applications Artículo de revista
En: ELCVIA: electronic letters on computer vision and image analysis, vol. 16, no 2, pp. 21–24, 2017.
Resumen | Enlaces | BibTeX | Etiquetas: COSFIRE Descriptor, Invariant Features, machine vision, Object Retrieval, Spermatozoa Classification
@article{fernandez-robles_recognition_2017,
title = {Recognition and retrieval of objects in diverse applications},
author = {Laura Fernández-Robles},
url = {https://www.raco.cat/index.php/ELCVIA/article/view/v16-n2-fernandez},
year = {2017},
date = {2017-01-01},
journal = {ELCVIA: electronic letters on computer vision and image analysis},
volume = {16},
number = {2},
pages = {21–24},
abstract = {This work focuses on object description and retrieval techniques applied to various real-world problems. It explores the classification of boar spermatozoa based on acrosome integrity using methods based on invariant local features. The paper also presents solutions for insert localization and recognition of broken inserts in milling heads, offering an automatic, in-process method for detection without interrupting machining operations. Additionally, it introduces a new descriptor, colour COSFIRE, for object retrieval in the context of the European project aimed at combating sexual exploitation of children.},
keywords = {COSFIRE Descriptor, Invariant Features, machine vision, Object Retrieval, Spermatozoa Classification},
pubstate = {published},
tppubtype = {article}
}
2015
Fernández-Robles, Laura; Alfonso-Cendón, Javier; Castejón, Manuel; García-Olalla, Óscar; Alegre, Enrique
DESARROLLO DE UNA APLICACIÓN DE RECUPERACIÓN DE OBJETOS Y ANÁLISIS DE LOS DESCRIPTORES LOCALES INVARIANTES UTILIZADOS Artículo de revista
En: 2015.
Resumen | Enlaces | BibTeX | Etiquetas: image analysis, Object Retrieval, SIFT
@article{fernandez-robles_desarrollo_2015,
title = {DESARROLLO DE UNA APLICACIÓN DE RECUPERACIÓN DE OBJETOS Y ANÁLISIS DE LOS DESCRIPTORES LOCALES INVARIANTES UTILIZADOS},
author = {Laura Fernández-Robles and Javier Alfonso-Cendón and Manuel Castejón and Óscar García-Olalla and Enrique Alegre},
url = {http://dspace.aeipro.com/xmlui/handle/123456789/725},
year = {2015},
date = {2015-01-01},
abstract = {This paper presents an application for the ASASEC project that helps retrieve objects from large image and video datasets related to child exploitation cases. By selecting regions of interest and using descriptors like SIFT and COSFIRE, the application finds similar objects across datasets. Experiments showed SIFT and COSFIRE outperformed SURF and HOG in retrieval accuracy.},
keywords = {image analysis, Object Retrieval, SIFT},
pubstate = {published},
tppubtype = {article}
}