[WEEK 1] Facade Parsing Using Deep Learning
Theme: Segmenting an image of a facade into predefined semantic categories
Team Members: Onur Cankur, furkan karababa, Javid Rajabov
This is our first blog about our project. It will be introduction of it and there will be brief explanation about our goal, datasets and related works.
Introduction:
We chose facade parsing as our research topic. It is about segmenting an image of a facade into predefined semantic categories. The categories will be like window, wall, sky, balcony, door, roof since we will segment facades of buildings.
This week, we analyzed the datasets and related works that we found last week. We have seen more than one way to accomplish our project by using different approaches. Papers are shared among us and each member in team read them. You can find brief explanations and examples of our datasets and related works below.
Datasets:
We are planning to use three datasets, Paris Art Deco Facades Dataset, eTRIMS image dataset, and ECP(Ecole Centrale Paris Facades Database) for this project.
Paris Art Deco Facades Dataset: This dataset contains images and labels of facades. In addition, the dataset consists seven labels such as door, shop, balcony, window, wall, sky, roof.
eTRIMS Dataset: We are planning to use 8-Class eTRIMS dataset. It contains classes such as window, door, sky, pavement, vegetation, car, road.
ECP Dataset: This dataset is also contains images and labels of facades and its labels door, shop, balcony, window, wall, sky, roof, outlier.
Related Works:
We are currently looking for different approaches to facade parsing and you can see some of the related papers that we are analyzing now.
