Implementation:Datajuicer Data juicer VideoFaceBlurMapper

Knowledge Sources	Datajuicer_Data_juicer
Domains	Data_Processing, Mapping
Last Updated	2026-02-14 16:00 GMT

Overview

Concrete tool for blurring faces detected in videos provided by Data-Juicer.

Description

VideoFaceBlurMapper is a mapper operator that detects and blurs faces in video frames to anonymize identities, supporting privacy-preserving data processing. It uses an OpenCV Haar cascade classifier (default: haarcascade_frontalface_alt.xml) to detect faces in each video frame, then applies a configurable blur kernel ("mean", "box", or "gaussian") with adjustable radius to the detected face regions, processing each frame and saving the result to the output directory.

Usage

Use when you need face anonymization for privacy and compliance requirements (such as GDPR) before using video data for training or distribution.

Code Reference

Source Location

• Repository: Datajuicer_Data_juicer
• File: data_juicer/ops/mapper/video_face_blur_mapper.py

Signature

@OPERATORS.register_module("video_face_blur_mapper")
class VideoFaceBlurMapper(Mapper):
    def __init__(self, cv_classifier: str = "", blur_type: str = "gaussian", radius: float = 2, save_dir: str = None, *args, **kwargs):

Import

from data_juicer.ops.mapper.video_face_blur_mapper import VideoFaceBlurMapper

I/O Contract

Inputs

Name	Type	Required	Description
cv_classifier	str	No	OpenCV classifier path for face detection (default: haarcascade_frontalface_alt.xml)
blur_type	str	No	Type of blur kernel: "mean", "box", or "gaussian" (default: "gaussian")
radius	float	No	Radius of the blur kernel (default: 2)
save_dir	str	No	Directory for generated video files; if not specified, saves alongside input files

Outputs

Name	Type	Description
samples	Dict	Transformed samples with face-blurred video file paths

Usage Examples

process:
  - video_face_blur_mapper:
      blur_type: "gaussian"
      radius: 3

Related Pages

• Environment:Datajuicer_Data_juicer_Python_Runtime_Environment