Please use this identifier to cite or link to this item:
http://hdl.handle.net/2445/133458
Title: | Fully convolutional architectures for multi-part body segmentation |
Author: | Borrego Carazo, Juan |
Director/Tutor: | Madadi, Meysam Escalera Guerrero, Sergio |
Keywords: | Xarxes neuronals (Informàtica) Aprenentatge automàtic Treballs de fi de màster Cos humà Processament digital d'imatges Neural networks (Computer science) Machine learning Master's theses Human body Digital image processing |
Issue Date: | 1-Sep-2018 |
Abstract: | [en] Since the appearance of the baseline Fully Convolutinal Network (FCN), convolution architectures usage has spread widely among Deep Neural Networks: from classification tasks to object tracking, they are found ubiquitously in the Deep Learning field. In this study, three different convolutional architectures are studied with regard its application to the semantic segmentation of the human body: ICNet, a different resolution cascade network, SegNet, a encoder-decoder network, and Stacked Hourglass, a specially purposed network for the human body. For this purpose, the SURREAL (Synthetic hUmans foR REAL tasks) dataset, which consists of synthetically rendered but realistic images of people, is used. As a result, is shown that the best performing network for this task is the Stacked Hourglass. Due to its continuous refinement of the output and the use of the full network for inference a 55.3% mIoU is achieved on the 24 body part dataset. |
Note: | Treballs finals del Màster de Fonaments de Ciència de Dades, Facultat de matemàtiques, Universitat de Barcelona, Any: 2018, Tutor: Meysam Madadi i Sergio Escalera Guerrero |
URI: | http://hdl.handle.net/2445/133458 |
Appears in Collections: | Màster Oficial - Fonaments de la Ciència de Dades |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
memoria.pdf | Memòria | 1.48 MB | Adobe PDF | View/Open |
This item is licensed under a
Creative Commons License