Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/81326
Title: El salto cualitativo de Deep Learning en problemas de clasificación
Author: López Bautista, María
Director: Morán Moreno, Jose Antonio  
Tutor: Martí Puig, Pere
Others: Universitat Oberta de Catalunya
Abstract: Biological processes are closely related to animals behavior, so their observation is getting more important. The most common method for this is video-recording at controlled situations for a long time. There are already algorithms that offer suitable results for tracking and detection of different individuals that cohabit in the same space. However, long-term tracking while maintaining the identity of animals is still a challenge. Besides, this can be more complicated when fish disappear from the camera's lens, cross each other or are blurred because of the quality of the image. Therefore, this master's thesis aims to propose a solution to this problem through machine learning methods application. These, with the tracking algorithms, will improve the information collected about the different individuals observed to draw conclusions about their behavior. The work focuses on the use of Deep Learning as the main algorithm for the classification and identification of fish. Later, it will be compared with other machine learning methods, concluding the qualitative leap that Deep Learning entails.
Keywords: SVM
deep learning
convolutional neural network
Document type: info:eu-repo/semantics/masterThesis
Issue Date: 10-Jun-2018
Publication license: http://creativecommons.org/licenses/by-nc-nd/3.0/es/  
Appears in Collections:Bachelor thesis, research projects, etc.

Files in This Item:
File Description SizeFormat 
mlopebauTFM0618.ppsxPresentación del TFM26,23 MBMicrosoft PowerpointView/Open
mlopebauTFM0618scripts.zipScripts desarrollados en el TFM47,49 kBWriter 6.0 documentsView/Open
mlopebauTFM0618memoria.pdfMemoria del TFM1,61 MBAdobe PDFThumbnail
View/Open
mlopebauTFM0618presentación.pdfPresentación en PPT del TFM770,61 kBAdobe PDFThumbnail
View/Open