Implementation:Datajuicer Data juicer ImageBlurMapper

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

Overview

Concrete tool for applying blur effects to images in the dataset provided by Data-Juicer.

Description

ImageBlurMapper is a mapper operator that blurs images using one of three supported blur types: mean, box, or Gaussian. The probability of an image being blurred is controlled by the p parameter. The blur effect is applied using a kernel with a specified radius via PIL's ImageFilter module. Blurred images are saved to a configurable output directory. The operator tracks original source file paths for provenance.

Usage

Use when you need to add controlled blur effects to training images for improving model robustness to image quality variations or for privacy-preserving data transformations.

Code Reference

Source Location

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

Signature

@OPERATORS.register_module("image_blur_mapper")
class ImageBlurMapper(Mapper):
    def __init__(self,
                 p: float = 0.2,
                 blur_type: str = "gaussian",
                 radius: float = 2,
                 save_dir: str = None,
                 *args, **kwargs):

Import

from data_juicer.ops.mapper.image_blur_mapper import ImageBlurMapper

I/O Contract

Inputs

Name	Type	Required	Description
p	float	No	Probability of the image being blurred, defaults to 0.2
blur_type	str	No	Type of blur kernel: mean, box, or gaussian; defaults to "gaussian"
radius	float	No	Radius of blur kernel, defaults to 2
save_dir	str	No	Directory to store generated image files; if not specified, saves in same directory as input

Outputs

Name	Type	Description
samples	Dict	Transformed samples with blurred image paths updated

Usage Examples

process:
  - image_blur_mapper:
      p: 0.2
      blur_type: "gaussian"
      radius: 2

Related Pages

• Environment:Datajuicer_Data_juicer_Python_Runtime_Environment