Add processing background summary to readme

[egrace479]

@egrace479 I agree with the suggested features. Here are some details on the software/model that's being used on the backend for these pipelines:

Model Inference

• Human-face identification: timesler/facenet-pytorch
• Museum specimen image filtering: CLIP ViT-L/14@336px
• Camera trap image filtering: Megadetector MDV6-yolov10-e

Apache Spark

• Citizen science image filtering pipeline: rule-based search
• PDQ hash-based image deduplication: faustomorales/pdqhash-python

Originally posted by @NetZissou in Imageomics/bioclip-ecosystem#8 (comment)