Dark Mode
Deep Learning Liver Analysis Dataset
Patient Health Records & Digital Health
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This dataset provides annotated ultrasound images of the liver, specifically curated to support the development of computer vision models. Its primary purpose is to aid in liver analysis, image segmentation, and the detection of liver diseases. The collection includes detailed annotations outlining the liver itself and any liver mass regions present. Images are classified into three distinct categories: benign, malignant, and normal cases, representing healthy livers, benign liver conditions, and malignant liver conditions, respectively. This makes the dataset suitable for training and evaluating models that differentiate between various liver states.
Columns
The accompanying dataset.csv file includes several columns detailing the characteristics of the image data:
- main_category / category: Identifies the primary classification of the ultrasound images, such as 'Benign', 'Malignant', or 'Other'. Benign and Malignant cases each account for 36% of the entries, with 'Other' making up 27%.
- sub_category_level_1 / sub dir: Indicates the type of data, whether it's 'segmentation' (73%) or a raw 'image' (27%).
- sub_category_level_2 / sub category of segmentations: Provides a further breakdown of segmentation types, including 'liver' (27%) and 'outline' (27%). Other categories account for 45% of entries, with a proportion of missing values.
- number_of_files / num of files: Specifies the count of files associated with each entry. The mean number of files is 258, with a standard deviation of 139, ranging from a minimum of 100 to a maximum of 435.
Distribution
The dataset is structured into three zip files, categorised by liver condition:
- Benign.zip: Contains ultrasound images classified as benign, with a file size of 16.9 MB.
- Malignant.zip: Contains ultrasound images classified as malignant, with a file size of 46.9 MB.
- Normal.zip: Contains ultrasound images of healthy livers, with a file size of 6.6 MB.
The dataset consists of ultrasound images with detailed annotations. While a specific total number of individual image files is not provided, the
dataset.csvfile offers insights into file counts per category.
Usage
This dataset is an ideal resource for various applications, including:
- Developing and evaluating deep learning models for the detection of liver diseases.
- Creating algorithms for the automated segmentation of liver and liver mass regions in ultrasound images.
- Conducting research in medical image analysis and computer-aided diagnosis systems.
- Serving as an educational tool for medical imaging studies and related academic fields.
Coverage
The dataset focuses exclusively on ultrasound images of the liver, providing insights into benign, malignant, and normal liver conditions. It was published on 2 November 2022. No specific geographical, temporal, or demographic scope has been provided.
License
Creative Commons Attribution 4.0 International
Who Can Use It
This dataset is particularly useful for:
- Computer Vision Developers: For building and refining models that analyse medical images.
- Medical Researchers: Those working on liver disease detection, diagnosis, and medical image analysis.
- Deep Learning Engineers: For training neural networks in medical image segmentation and classification tasks.
- Educators and Students: As a practical resource for learning about medical imaging and AI applications in healthcare.
Dataset Name Suggestions
- Liver Ultrasound Imaging Dataset
- Annotated Liver Scan Dataset for AI
- Medical Liver Pathology Ultrasound Data
- Deep Learning Liver Analysis Dataset
- Ultrasound Liver Disease Classification Set
Attributes
Original Data Source: Deep Learning Liver Analysis Dataset
QUALITY