Mobiusi's picture
initial commit
fbb0c1a verified
metadata
tags:
  - Image Classification
  - Object Detection
  - Industrial Inspection
  - Quality Control
  - Defect Detection
license: cc-by-nc-sa-4.0
task_categories:
  - image-classification
language:
  - en
pretty_name: Mold Template Defect Recognition Dataset
size_categories:
  - 1B<n<10B

Mold Template Defect Recognition Dataset

In the industrial sector, mold quality inspection is crucial for ensuring product integrity but faces challenges such as inconsistent defect detection and high rates of false negatives. Existing solutions often rely on manual inspection, which is time-consuming and prone to human error, leading to inefficiencies. This dataset aims to address these challenges by providing a comprehensive collection of labeled images that enhance machine learning models for accurate defect recognition in mold templates. Data was collected using high-resolution cameras in controlled industrial environments, ensuring clarity and consistency. Quality control measures included multiple rounds of labeling, consistency checks among annotators, and expert reviews to maintain high accuracy. The images are stored in JPG format, organized in directories by defect type, facilitating easy access and processing.

Technical Specifications

Field Type Description
file_name string File name
quality string Resolution
defect_type string The types of defects appearing on the mold, such as cracks and pores.
defect_location string A detailed description of the defect's specific location on the mold.
severity_level integer The severity level of the defect as defined by the standards.
bounding_box_coordinates string Coordinates of the bounding box that marks the defect in the target detection task.
surface_condition string The overall condition of the mold surface, including the presence of oxidation or wear.
light_condition string The lighting conditions when capturing the image, such as natural light or artificial light.
image_quality string The quality assessment of the image, such as clarity or blurriness.

Compliance Statement

Authorization Type CC-BY-NC-SA 4.0 (Attribution–NonCommercial–ShareAlike)
Commercial Use Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization No PII, no real company names, simulated scenarios follow industry standards
Compliance System Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

Source & Contact

If you need more dataset details, please visit Mobiusi. or contact us via contact@mobiusi.com