For a Halloween-themed experiment, we used machine learning to help analyze the scene settings and content in the 1978 movie “Halloween.” We used TensorFlow img2txt to generate descriptions for each frame and then ran python scripts to pull frames based on word descriptions. Using the specific content we could get specific percentages and frame amounts for quantitative data surrounding female stereotypes.
Here is a link to the data visualization of the frames.
Time: 3 weeks
Technology: Python, Tensorflow- im2txt
Goal: Analyze gender to reveal patterns in the portrayal of women in film using machine learning
Credits: Aarati Akkapeddi, Tresson Canley, Lauren Malkani