C2D2 Coverage Prediction - 3.5 English

Vitis AI Library User Guide (UG1354)

Document ID
Release Date
3.5 English

Colonoscopy Coverage Deficiency via Depth algorithm, or C2D2, is a machine learning-based approach for improving colonoscopy coverage. The C2D2 network is a cascading structure. The inputs are 300 serialized gray images and the output is coverage. The C2D2_Lite_0_pt model is responsible for extracting the features of each image and the C2D2_Lite_1_pt model predicts a coverage value by inputting the characteristics of 300 pictures.

The following table lists the C2D2 Coverage Prediction models supported by the Vitis AI Library.

Table 1. C2D2 Model
No Model Name Framework
1 C2D2_Lite_0_pt PyTorch
2 C2D2_Lite_1_pt PyTorch