Painting Analysis

Project Proposals P2

CS 585 Fall2018
Kaiyuan Fan
kaikang Zhu

Project Problem Definition

Help art historians to analyze paintings. Prolific painters may change their style throughout their lifetime. We are going to pick an artist with varied paint styles during his or her lifetime. A good example provided is the Italian artist Tiziano Vecelli (Titian). We are going to process the painterís color scheme to predict the painting age, recognize objects in paintings, analyze the location of faces, etc. We have divided our tasks into the following parts:

  1. Detect, count and label the people in the painting. Analysis if they are naked or not.
  2. Training with classified dataset. Then given a test painting, we should recognize it's color scheme and estimate the year of production.
Painting Year Classification:
Early Years (1500-1529)
Middle Years (1530-1550) - In 1525 he married Cecilia (who tragically died in 1530). Titianís meeting with the Holy Roman Emperor Charles V in Bologna in 1530 would be a determining event in his life.
Late Years (1550+) - Starting from the late 1550s, Titian developed a much freer use of the brush and a less descriptive representation of reality.

Desired Input and Output

Inputs: Titian's paintings
Example input and output:example input => Early Year(1500-1529)/ People: 5/ Naked: 1 example output

Related Paper and Algorithms

  • Machine Learning Algorithms That Can Analyze Works of Art
  • Using Machine Learning for Identification of Art Paintings
  • Detect People in Cubist Art

  • What we have done so far

    We currently have prepared and renamed the dataset we are going to use. We have around 200 paintings from Titian.
    We haven't made much significant implementation yet. But we have a great idea of how to start doing the project. We will have a training dataset to calculate percentage of red, blue and green pixels in different classified years. After that, we will process with the test painting dataset, mapping them to the most closing class.