LDPolypVideo - sporedata/researchdesigneR GitHub Wiki
General description
The LDPolypVideo Benchmark is a large public dataset and evaluation benchmark primarily used in the field of medical imaging, particularly for the detection, segmentation, and classification of polyps during colonoscopy procedures. The LDPolypVideo Benchmark consists of 160 videos with 40,266 frames, and plays a crucial role in the development of computer-aided detection (CAD) and deep learning algorithms aimed at improving the accuracy of polyp detection and segmentation in real-time, which is critical for early diagnosis of colorectal cancer.
Data Categories
- Polyp types.
- Video data.
Limitations
- Motion Blur: Due to camera movement during the procedure, frames may experience blur, which challenges the performance of detection algorithms.
- Lighting Conditions: Colonoscopy videos often have inconsistent lighting due to the reflective nature of mucosal tissue, shadowing, and camera movement. This adds complexity to the task of detecting and segmenting polyps accurately.
- Variability in Polyps: Polyps vary greatly in size, shape, color, and texture, making it challenging for detection algorithms to generalize across different polyp types.
- Occlusion and Obstructions: Polyps can sometimes be partially obscured by folds in the colon or debris, requiring robust models that can detect partially visible polyps.
Related publications
Data access
For more information on the LDPolypVideo Benchmark dataset, visit https://github.com/dashishi/LDPolypVideo-Benchmark