metadata

language:
  - en
license: mit
multilinguality:
  - monolingual
size_categories:
  - 1K<n<10K
source_datasets:
  - original
task_categories:
  - image-classification
  - object-detection
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
dataset_info:
  features:
    - name: id
      dtype: string
    - name: image
      dtype: string
    - name: target
      dtype: string
    - name: instrument
      dtype: string
    - name: filter
      dtype: string
    - name: date_obs
      dtype: string
    - name: exptime
      dtype: 'null'
    - name: ra
      dtype: 'null'
    - name: dec
      dtype: 'null'
    - name: program
      dtype: string
    - name: image_path
      dtype: string
    - name: width
      dtype: int64
    - name: height
      dtype: int64
    - name: total_pixels
      dtype: int64
    - name: mean_intensity
      dtype: float64
    - name: std_intensity
      dtype: float64
    - name: min_intensity
      dtype: float64
    - name: max_intensity
      dtype: float64
    - name: median_intensity
      dtype: float64
    - name: skewness
      dtype: float64
    - name: kurtosis
      dtype: float64
    - name: dynamic_range
      dtype: float64
    - name: noise_level
      dtype: float64
    - name: noise_std
      dtype: float64
    - name: signal_to_noise
      dtype: float64
    - name: saturated_pixels
      dtype: int64
    - name: saturation_percentage
      dtype: float64
    - name: is_saturated
      dtype: bool
    - name: cosmic_rays
      list:
        list: int64
    - name: hot_pixels
      list:
        list: int64
    - name: bad_pixels
      list: 'null'
    - name: artifact_count
      dtype: int64
    - name: quality_score
      dtype: float64
  splits:
    - name: train
      num_bytes: 18984918
      num_examples: 2709
  download_size: 3821834
  dataset_size: 18984918

JWST Quality Analysis Dataset

Overview

This dataset contains comprehensive quality analysis for 2,709 JWST (James Webb Space Telescope) NIRCam images from the MAST archive. Each image has been automatically analyzed for quality metrics, artifact detection, and noise characteristics.

Dataset Information

Size: 2,709 images
Format: JSONL (JSON Lines)
Source: JWST NIRCam observations from MAST
Targets: M16, NGC 3132, NGC 3324, SMACS 0723, Stephan's Quintet

Quality Metrics

Each image includes:

Basic Statistics

mean_intensity, std_intensity, min_intensity, max_intensity
median_intensity, skewness, kurtosis, dynamic_range

Noise Analysis

noise_level, noise_std, signal_to_noise

Saturation Analysis

saturated_pixels, saturation_percentage, is_saturated

Artifact Detection

cosmic_rays: List of cosmic ray locations [x, y, area]
hot_pixels: List of hot pixel locations [x, y, area]
bad_pixels: List of bad pixel locations [x, y, area]
artifact_count: Total number of artifacts

Quality Assessment

quality_score: Overall quality score (1-10 scale)

Use Cases

For Researchers

Quality Screening: Filter images by quality score for analysis
Artifact Cataloging: Identify and locate artifacts for cleaning
Statistical Analysis: Study image quality across different targets/filters
Quality Benchmarking: Compare quality across different observations

For Machine Learning

Training Data: Train quality assessment models
Feature Engineering: Use quality metrics as features
Validation: Quality scores for model evaluation

Methodology

The quality analysis was performed using:

OpenCV for image processing and artifact detection
NumPy/SciPy for statistical analysis
Parallel processing for efficient analysis of large datasets

Quality scores are calculated based on:

Signal-to-noise ratios
Saturation levels
Artifact counts
Dynamic range

Dataset Statistics

Mean Quality Score: 8.39/10
Quality Score Range: 6.0 - 10.0
Images with Artifacts: All images contain some artifacts (typical for astronomical data)
Saturated Images: 0 (no significant saturation detected)

Citation

If you use this dataset in your research, please cite:

@dataset{jwst_quality_analysis_2024,
  title={JWST Quality Analysis Dataset},
  author={Your Name},
  year={2024},
  url={https://huggingface.co/datasets/norbertm/jwst-quality-analysis-dataset}
}

License

This dataset is provided for research purposes. Please refer to the original JWST data usage policies from MAST.

Contact

For questions or feedback about this dataset, please open an issue on the Hugging Face repository.