README.md

metadata

license: cc-by-4.0
task_categories:
  - visual-question-answering
language:
  - en
tags:
  - manufacturing
  - additive-manufacturing
  - directed-energy-deposition
  - ded
  - thermal-imaging
  - vqa
pretty_name: DED Thermal Visual Question Answering (DD4)
size_categories:
  - 1K<n<10K
configs:
  - config_name: default
    data_files:
      - split: train
        path: train.parquet
      - split: test
        path: test.parquet

DED-Thermal VQA Dataset

This dataset is designed for Visual Question Answering (VQA) in the domain of advanced manufacturing, specifically focusing on Directed Energy Deposition (DED) additive manufacturing processes. It pairs in-situ infra-red (IR) thermal monitoring sequences/images with structured question-answer pairs to facilitate the development of Vision-Language Models (VLMs) capable of understanding physical phenomenon and thermal history in 3D printing.

---

Dataset Summary

During the Directed Energy Deposition (DED) process, real-time thermal behavior (e.g., melt pool dynamics, cooling rates, and thermal gradients) is critical to the structural integrity and quality of the final printed components.

dataset provides a benchmark for evaluating multimodal AI agents on their ability to interpret, reason, and answer questions about physical manufacturing dynamics based on thermal visualization data.

• Task: Visual Question Answering (VQA)
• Domain: Additive Manufacturing / Directed Energy Deposition (DED)
• Modality: Multi-modal (Thermal Imaging + Natural Language)
• Format: Parquet (compatible with Hugging Face datasets library)

---

Dataset Structure

The directory structure of this dataset is as follows:

vqa_ded_thermal/
├── train.parquet          # Training split containing VQA pairs and image features/references
├── test.parquet           # Testing split for model evaluation
├── dataset_report.json    # Statistical overview and structural metadata of the dataset
└── README.md              # Dataset card and documentation