Back
Texte intégral
Bachelor thesis

A machine learning approach for digital image restoration

SONAR|HES-SO

    2017

1 ressource en ligne

Mémoire de bachelor: Haute Ecole de Gestion & Tourisme, 2017

Image detection
Random forest
Neural network
Hough transform
Contour filtering
English This paper illustrates the process of image restoration in the sense of detecting images within a scanned document such as a photo album or scrapbook. The primary use case of this research is to accelerate the cropping process for the employees of Cinetis, a company based in Martigny, Switzerland that specializes in the digitalization of old media formats. In this paper, we will first summarize the state of the art in this field of research. This will include explanations of various techniques and algorithms involved with feature and document detection used by various digital companies.
Language
  • English
Classification
Computer science and technology
Notes
  • Haute Ecole de Gestion Valais
  • Informatique de gestion - Wirtschaftsinformatik
  • hesso:hegvs
License
License undefined
Identifiers
  • RERO DOC 309509
  • RERO R008808165
Persistent URL
https://folia.unifr.ch/global/documents/316735
Statistics
Document views: 141 File downloads:
  • MAYORAZ_CALIXTE_2017.pdf: 303