Enhancing Object Segmentation via Few-Shot Learning with Limited Annotated Data

Simple item page
dc.contributor.authorGarcía-Aguilar, Iván
dc.contributor.authorJafri, Syed Ali Haider
dc.contributor.authorElizondo, David
dc.contributor.authorCalderón, Saul
dc.contributor.authorGreenfield, Sarah
dc.contributor.authorLuque-Baena, Rafael M.
dc.date.acceptance2024-07-19
dc.date.accessioned2025-01-14T16:46:24Z
dc.date.available2025-01-14T16:46:24Z
dc.date.issued2024-11-20
dc.description.abstractSignificant advancements in machine learning in recent years have revolutionized multiple sectors. The Segment-Anything Model (SAM) is a notable example of state-of-the-art image segmentation. Despite claims of zero-shot generalization, SAM exhibits limitations in specific scenarios like medical mammography images. SAM generates three segmentation masks per image to address this and recommends selecting the one with the highest confidence score. However, this is not always the optimal choice. This paper introduces a system that extends SAM’s segmentation capabilities by automatically selecting the correct mask, leveraging few-shot learning methods and an Out-of-Distribution threshold strategy. Several backbones were subjected to experimentation, highlighting the relationship between the support set size and the model’s accuracy.
dc.funderNo external funder
dc.identifier.citationGarcía-Aguilar, I., Ali Haider Jafri, S., Elizondo, D., Calderón, S., Greenfield, S., M. Luque-Baena, R. (2025) Enhancing Object Segmentation via Few-Shot Learning with Limited Annotated Data. In: Quintián, H., et al. The 19th International Conference on Soft Computing Models in Industrial and Environmental Applications SOCO 2024. SOCO 2024. Lecture Notes in Networks and Systems, vol 889. Springer, Cham
dc.identifier.doihttps://doi.org/10.1007/978-3-031-75010-6_4
dc.identifier.isbn9783031750090
dc.identifier.isbn9783031750106
dc.identifier.issn2367-3370
dc.identifier.issn2367-3389
dc.identifier.urihttps://hdl.handle.net/2086/24697
dc.language.isoen
dc.peerreviewedYes
dc.publisherSpringer Nature
dc.relation.ispartofLecture Notes in Networks and Systems
dc.relation.ispartofThe 19th International Conference on Soft Computing Models in Industrial and Environmental Applications SOCO 2024
dc.researchinstitute.instituteInstitute of Digital Research, Communication and Responsible Innovation
dc.titleEnhancing Object Segmentation via Few-Shot Learning with Limited Annotated Data
dc.typeBook chapter

Files

License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
4.2 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

Leicester Media School