Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/63945
Title: Machine learning regularity representation from biological patterns: a case study in a Drosophila neurodegenerative model
Author: Díez Hermano, Sergio
Tutor: Vegas Lozano, Esteban
Others: Universitat Oberta de Catalunya
Abstract: This work presents a fully automated classification pipeline of bright-field images based on HOG descriptors and machine learning techniques. An initial ROI extraction is performed applying TopHat morphological kernel and Euclidean distance to centroid thesholding. Image classification algorithms are trained on these ROIs (SVM, Decision Trees, Random Forest, CNN) and their performance is evaluated on independent, unseen datasets. HOG + gaussian kernel SVM (0.97 accuracy and 0.98 AUC) and fine-tune pre-trained CNN (0.98 accuracy and 0.99 AUC) yielded the best results overall.
Keywords: Drosophila melanogaster
machine learning
algorithms
Document type: info:eu-repo/semantics/masterThesis
Issue Date: May-2017
Publication license: http://creativecommons.org/licenses/by-nc-sa/3.0/es/  
Appears in Collections:Trabajos finales de carrera, trabajos de investigación, etc.

Files in This Item:
File Description SizeFormat 
sdiezhTFM0617memory.pdfMemoria del trabajo fin de máster2,53 MBAdobe PDFThumbnail
View/Open