rf100-vl-datasets / liver-disease /README.dataset.txt
gatilin's picture
Upload folder using huggingface_hub
e7852cb verified
# liver-disease-jyvvu-fash > 2025-03-13 9:26pm
https://universe.roboflow.com/rf-100-vl/liver-disease-jyvvu-fash
Provided by a Roboflow user
License: MIT
# Overview
- [Introduction](#introduction)
- [Object Classes](#object-classes)
- [Ballooning](#ballooning)
- [Fibrosis](#fibrosis)
- [Inflammation](#inflammation)
- [Steatosis](#steatosis)
# Introduction
This dataset focuses on identifying pathological features in liver tissue samples related to liver disease. The objective is to annotate images with specific liver disease conditions: ballooning, fibrosis, inflammation, and steatosis. Each class represents a distinct morphological change within the tissue that is important for diagnosing and understanding liver disease progression.
- **Ballooning**: Liver cells are enlarged and swolen.
- **Fibrosis**: Liver damage due to scar tissue in the liver
- **Inflammation**: A cluster of darker, denser cells.
- **Steatosis**: Liver damage due to a buildup of fat.
# Object Classes
## Ballooning
### Description
Ballooning refers to enlarged liver cells, often appearing as large, round, clear spaces within the tissue. These spaces usually have a well-defined perimeter and a relatively homogeneous texture inside.
### Instructions
- Annotate regions that exhibit large, round, clear spaces with distinct boundaries.
- Do not include irregular shapes or areas that do not have clear perimeters. Focus only on the well-defined circular regions.
## Fibrosis
### Description
Fibrosis is characterized by a dense, web-like pattern of tissue, often in thread-like strands. These strands are usually depicted as elongated, interwoven structures in the tissue. These tissues are often a darker shade of blue compared to surrounding tissue.
### Instructions
- Outline the web-like, elongated strands within the tissue.
- Avoid labeling isolated or fragmented areas that do not contribute to the fibrous network. Focus on continuous or interconnected strands.
## Inflammation
### Description
Inflammation is depicted by clusters of darker, denser cellular areas. These clusters often appear as small, dense aggregations and may be irregular in shape.
### Instructions
- Identify small, dense clusters that appear darker compared to surrounding tissue.
- Do not annotate dispersed cells that do not form a compact group. Select regions where cellular density is markedly higher.
## Steatosis
### Description
Steatosis is indicated by large droplets of fat within the liver cells, often seen as large round areas with distinct borders that are lighter in appearance.
### Instructions
- Mark large, filled round areas with distinct edges, focusing on rounded, light regions.
- Avoid marking smaller or misshapen areas that do not clearly represent fat droplets. Prioritize obvious, sizeable droplets.