33 lines
1.3 KiB
Plaintext
33 lines
1.3 KiB
Plaintext
|
You want to build a custom AI-powered system to automatically generate keyword tags for agricultural stock photos.
|
|||
|
|
|
|||
|
|
You want the system to help eliminate your current manual keyword tagging process, which is currently handled by an assistant and takes about 10 hours/month.
|
|||
|
|
|
|||
|
|
You need to process 1,000 photos per month, in batches of 500, and this number may scale up over time (possibly doubling in 3 years).
|
|||
|
|
|
|||
|
|
You want the system to generate 5 to 10 high-quality keywords per image, with a focus on agricultural relevance.
|
|||
|
|
|
|||
|
|
You want to be able to train the AI using your current keyword-tagged photo dataset, which contains about 30,000 photos.
|
|||
|
|
|
|||
|
|
The system must differentiate subtle ag-specific concepts, such as:
|
|||
|
|
|
|||
|
|
Farmer vs. rancher
|
|||
|
|
|
|||
|
|
Dairy farmer vs. rancher
|
|||
|
|
|
|||
|
|
Chicken farmer (not rancher)
|
|||
|
|
|
|||
|
|
Male vs. female farmers (for diversity tagging)
|
|||
|
|
|
|||
|
|
You want the system to optionally generate a descriptive product title like: “Farmer and son walking in cornfield.”
|
|||
|
|
|
|||
|
|
If location metadata is available in the image file, you want the system to extract and use that data as a keyword (e.g., “Iowa”).
|
|||
|
|
|
|||
|
|
You want the final output in CSV format, with each photo’s file name matched to its:
|
|||
|
|
|
|||
|
|
Human-entered keywords (for comparison, if needed)
|
|||
|
|
|
|||
|
|
AI-generated keywords
|
|||
|
|
|
|||
|
|
AI-generated title (if available)
|
|||
|
|
|
|||
|
|
Location (if available)
|