Face De-Identification: A Domain-Centric Survey from Capture to Processing

Researchers released a domain-centric survey of face de-identification methods spanning physical, sensor, and digital stages of the imaging pipeline.

The paper organizes techniques across capture and post-processing, giving privacy and computer vision communities a unified view of the field.

It also reviews current evaluation practices and notes that benchmarking remains inconsistent across methods and domains.

A key takeaway is the need for standard protocols and benchmarks to compare face de-identification approaches more fairly.