I noted that the US TV advertising revenue decline of -11.6% in 2020 (PwC) motivates the shift toward visual‑driven promotion such as movie‑poster genre prediction.
I chose Labellerr’s automated annotation platform to pre‑label movie posters, avoiding manual labeling and enabling rapid genre‑prediction model training.
I explored using AI to predict movie genres directly from poster images, aiming to help studios place ads more effectively after seeing a downturn in TV ad spend.
According to PwC, there is a -11.6% year‑on‑year decline in TV advertising revenue for the year 2020 in the US.
The workflow I envision uses Labellerr’s automated annotation to pre‑label poster images, then trains a computer‑vision model to output genre predictions that can be fed into ad‑placement pipelines.