Sif92/ultrasound-organ-dataset

Ultrasound Organ Dataset

Overview

The Ultrasound Organ Dataset is a curated collection of 5,005 ultrasound (US) images of abdominal organs collected from 563 patients at MH Samorita Medical College and Hospital, Dhaka, Bangladesh.

The dataset was designed for deep learning research in medical imaging, supporting tasks such as:

• Organ classification
• Anomaly detection
• Semi-supervised learning
• Medical AI research for resource-constrained healthcare environments

Ultrasound imaging is widely used because it is non-invasive, inexpensive, and free of radiation risk, making it particularly valuable for global healthcare applications.

---

Organ Classes

The dataset contains images from 10 abdominal organ categories:

• Abdominal Aorta
• Gallbladder
• Hepatic Vein
• Kidneys
• Liver
• Ovaries
• Pancreas
• Portal Vein
• Spleen
• Urinary System (Urinary Bladder, Prostate, and Uterus)

---

Dataset Structure

Radiologist 1

organ_classification_1

• 2,784 labeled images across the 10 organs.

anomaly_detection_1

• Normal: 2,014 images
• Abnormal: 799 images

organ_classification + anomaly_detection
Hybrid dataset combining organ classification and anomaly detection tasks, including an additional ascites abnormal class.

---

Radiologist 2

organ_classification_2

• 1,293 labeled organ images.

anomaly_detection_2

• Normal: 656 images
• Abnormal: 269 images

Used for semi-supervised anomaly detection experiments.

---

Patient_Wise

Contains 170 patient records with diagnostic metadata (xlsx and text file).

---

Data Collection

Images were collected during routine clinical examinations conducted by two radiologists at MH Samorita Medical College and Hospital.

All images were carefully organized and labeled, with systematic folder structures designed to facilitate reproducible machine learning experiments.

---

Ethics Approval

This dataset was collected under formal approval from the Institutional Ethical Review Board (IERB) of MH Samorita Medical College and Hospital, Dhaka, Bangladesh.

All images were fully anonymized and de-identified before releasing to ensure patient privacy and compliance with ethical research standards.

---

Dataset Access

📥 Official dataset download

https://scholarsjunction.msstate.edu/research-data/5/

---

Potential Applications

• Medical image classification
• Ultrasound image analysis
• Deep learning for healthcare
• Semi-supervised and unsupervised anomaly detection
• Computer-aided diagnosis systems

---

License

This dataset is released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.

Users must cite this dataset when using it in publications, research, or derivative works.

---

Acknowledgments

This dataset was developed as part of research conducted at Mississippi State University under the supervision of Dr. John E. Ball.

The author thanks:

• MH Samorita Medical College and Hospital
• The participating radiologists and sonographers
• Clinical staff who assisted in ultrasound data collection

Special thanks to Dr. Md. Enayet Karim for coordinating the acquisition of ultrasound images and metadata.

---

Citation

If you use this dataset in your research, please cite the associated dataset or publication.