Jump to content

Connect SuperML | Leeroopedia MCP: Equip your AI agents with best practices, code verification, and debugging knowledge. Powered by Leeroo — building Organizational Superintelligence. Contact us at founders@leeroo.com.

Implementation:Datajuicer Data juicer ImageSegmentMapper

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

Overview

Concrete tool for performing segment-anything on images using FastSAM provided by Data-Juicer.

Description

ImageSegmentMapper is a mapper operator that uses a FastSAM model (FastSAM-x.pt or FastSAM-s.pt) via the ultralytics library to detect and segment objects in images, returning their bounding boxes. It processes each image with configurable resolution, confidence threshold, and IoU threshold. Bounding boxes in xywh format are extracted from retina masks output and stored in the sample's metadata under bbox_tag as N x M x 4 arrays (N images, M boxes, 4 coordinates). Requires CUDA acceleration and approximately 800MB of GPU memory.

Usage

Use when you need region-level annotations for images, understanding spatial composition, detecting objects of interest, and supporting downstream image analysis tasks in multimodal data pipelines.

Code Reference

Source Location

Signature

@OPERATORS.register_module("image_segment_mapper")
class ImageSegmentMapper(Mapper):
    def __init__(self,
                 imgsz=1024,
                 conf=0.05,
                 iou=0.5,
                 model_path="FastSAM-x.pt",
                 *args, **kwargs):

Import

from data_juicer.ops.mapper.image_segment_mapper import ImageSegmentMapper

I/O Contract

Inputs

Name Type Required Description
imgsz int No Resolution for image resizing, defaults to 1024
conf float No Confidence score threshold, defaults to 0.05
iou float No IoU (Intersection over Union) score threshold, defaults to 0.5
model_path str No Path to FastSAM model: FastSAM-x.pt or FastSAM-s.pt; defaults to "FastSAM-x.pt"

Outputs

Name Type Description
samples Dict Transformed samples with bounding boxes stored in meta field under bbox_tag

Usage Examples

process:
  - image_segment_mapper:
      imgsz: 1024
      conf: 0.05
      iou: 0.5
      model_path: "FastSAM-x.pt"

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.
Principle
Implementation
Heuristic
Environment