This master thesis will be focused on the classification and detection of specific types of decay recognizable within images of historic buildings. The careful study on the related literature to collect the most important papers on the topic, the work is focused on historic building. A Deep Learning/Machine Vision-based system able to detect and classify with a certain confidence two categories of decay, Mask R-CNN has been chosen to perform a classification of decay. The prime step of thesis is dataset analysis to extract statistics, a step of dataset preparation and preprocessing like resizing images, creatin image annotations in COCO format is useful as an input to the deep learning.

APPROCCI BASATI SU DEEP LEARNING PER IL RICONOSCIMENTO DEL DEGRADO IN IMMAGINI RIGUARDANTI EDIFICI STORICI

CHINTHAKINDI, MANISH
2020/2021

Abstract

This master thesis will be focused on the classification and detection of specific types of decay recognizable within images of historic buildings. The careful study on the related literature to collect the most important papers on the topic, the work is focused on historic building. A Deep Learning/Machine Vision-based system able to detect and classify with a certain confidence two categories of decay, Mask R-CNN has been chosen to perform a classification of decay. The prime step of thesis is dataset analysis to extract statistics, a step of dataset preparation and preprocessing like resizing images, creatin image annotations in COCO format is useful as an input to the deep learning.
2020
A DEEP LEARNING-BASED APPROACH FOR DECAY RECOGNITION INTO IMAGES ABOUT HISTORIC BUILDINGS.
File in questo prodotto:
Fulltext non presente su UNITESI – Verifica la disponibilità della tesi in formato cartaceo interrogando il Catalogo di Ateneo

I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.14253/91012