Image comparing software Part One: Let’s look at some art!

The idea behind this project is the disparities that exist in the composition of color used in different styles of painting.

These differences could firstly be used to identify the style, and secondly used to cross-examine art of the same style but through different artists or different pieces.

Impressionism

These paintings are Impressionist and although they don’t look extremely similar, they are of the same style and all use a ton of color.

Baroque

These are baroque-era paintings and they have a lot of darker hues to them.

Cubism

The final category I decided to include are Cubist paintings which are mostly of Picasso. These are extremely colorful and mosaic.

In the next post, you’ll be able to see code snippets and analysis of the image pixel data!

Image Comparing Software, Part Two: Individual results, modules, and methods.

Extracting Pixel Data The way I went about analyzing our image data is by using PIL, or the Python Imaging Library. It allows us to grab the pixel data for each individual painting in our dataset. This would allow me to get the RGB data, and convert it into its closest available X11 color which…
Read more…

July 30, 2018

Bryan Alexis

Journal, Pixel Analysis Journals

Art, Image, Picasso, Python

Statement of Purpose

Please describe the project you would like to conduct in terms that can be understood by a non-expert audience.

My main project goal is in testing, through quantitative and qualitative research, my hypothesis that a person’s self-curated online environment can affect or predict their developmental growth. In the field of human development, I found it interesting that environments are able to shape the kinds of interactions humans face regularly, and that these interactions can be used to interpret a person’s development. The dawning of the internet has introduced a new vector through which we are being influenced to make real-world decisions, and I think it may be important to research the effects that it will have. I hypothesize that the internet’s influences are happening covertly a majority of the time, and that there exists little possibility for an average person to understand how they are being influenced.

I believe that data science is a field which is helpful in this regard because it will allow me to incorporate human development research in the real world and apply it in an online, data-centered environment. A significant aspect of human development research is that they conduct their findings using field research in specific environments (e.g. a nursery, playground, home, work-environment). It is important to track human environments in this way because the data gathered is contextual. The data gathered in a nursery, for example, has more to do with a child’s development than an adult. One environment I believe is missing from the study of human development is the online environment.

My main hypothesis is that a human’s interaction with their “personally curated” online environment has the ability to impact their future development. One such example is the invisible effect that an online identity can incur in the real-world, or how your real-life upbringing can differentiate your online persona. My project aims to capture what is missing in human development using HCI. I truly believe the intersection of these two fields has been largely ignored but which can explain many uniquely human issues which have arisen due to new technologies.

My project will be applying the fields of Human Development and Cognitive Science with data science techniques. I wish to develop research of an online environment (e.g. YouTube, Reddit, Twitter, Facebook). The “environment” will serve to inform my analysis of unique markers in Human Development: Such as goals, attitudes, and other developmental effects. It will also provide me with a valuable perspective from where to begin my research. As an example, a public forum, such as Reddit, will contain less personal information but may contain valuable unique perspectives/markers that may hint at trends in YouTube “personalities”. My analysis will be based on the body of knowledge present within human development. This knowledge would help in categorizing the “markers” present within these datasets/environments which may hint to a broader trend among humans.

Currently, I aim to utilize simple linear regression and statistical analysis. However, since identifying markers in data is something only humans can do, machine learning and sentiment analysis might also be useful for our purposes.

Independance

This Scholarship is meant to support your independent project, which should be supervised and mentored by a faculty member, but which should be your own work and responsibility, rather than something that mostly “belongs” to someone else in your research group.