UChicago-Computational-Content-Analysis / Readings-Responses-2024-Winter

1 stars 0 forks source link

9. Large Multi-Modal Models (LMMMs) to Incorporate Images, Art & Video - [E4]. Sgourev, S. V., Aadland, E., & Formilan, G. 2023. #3

Open lkcao opened 7 months ago

lkcao commented 7 months ago

Post questions here for this week's exemplary readings:

  1. Sgourev, S. V., Aadland, E., & Formilan, G. 2023. “Relations in Aesthetic Space: How Color Enables Market Positioning.” Administrative Science Quarterly, 68(1), 146-185.
XiaotongCui commented 6 months ago

Very fun research! I have a question about deciding whether a band is a pure Black Matel Brand or not. Will there be a possibility that the brand will change its style later on? In this case some bands may be pure at first and not pure later. I wonder does this research considers this point.

chenyt16 commented 6 months ago

This is a really interesting study. I believe identifying the optimal sample for generalizing this study to the commercial market might pose a challenge. One contributing factor is the consistent use of color for corporate identity, contrasting with the fluctuating colors employed in the campaign's key visual.

sborislo commented 6 months ago

I think it’s super cool to study color as its components, and I wonder whether there are other ways to break images down into their components in impactful ways. For instance, many advertisements these days are becoming “busier,” or showing more differentiable “sub-images” in the advertisement at once. Is there a way to break advertisements or media content down into a measure of “business” like they break down color?

muhua-h commented 6 months ago

What was the rationale behind choosing k-means over other clustering algorithms, such as DBSCAN or hierarchical clustering, especially considering the potential for noise in the data?

donatellafelice commented 6 months ago

the authors write: "But color does more than just materialize internal emotional states—it is a social technology and an organizing tool that accomplishes social work as a medium of relationships between subject and object... " i was trying to think of other organizational tools from the visual or audio world that could be explored for social science research. for example, specific shapes are deployed in art work to suggest 'hidden' messages, sometimes paintings in certain periods would depict objects that took on double meanings, and there is any number of chord progressions that have specific associations and resolutions. i wonder which would lend themselves best to computational anlaysis?

runlinw0525 commented 6 months ago

This study is very interesting. I wonder that how, then, can future research improve our understanding of how brightness and contrast interact in our perception of colors, particularly given the challenge of accurately modeling this complexity in visual studies?

QIXIN-LIN commented 6 months ago

Exploring color usage and its significance is fascinating! However, the interpretation of color varies widely among individuals - often, even the creators of album covers may not deeply consider their color choices. Furthermore, the meaning attributed to colors is constantly evolving. This variability raises questions about the reliability of assigning specific meanings to colors. Additionally, it prompts curiosity about other visual elements that might be studied similarly. How might these elements convey varied meanings, and how can we understand their impact in a broader context?

YucanLei commented 5 months ago

I am curious if the color usage could be influential for the games in fact, relevant to my project, I am wondering if the color usage could be influential to the audiences of the games, thus influencing the reviews the game could receive.