List your company
ResearchFormerly WizWand

Pathological Classification (6 classes) on ICBHI challenge dataset 2017 (test)

99.05AccuracyProposed model
97.54297.933598.32598.7165Sep 8, 2020
Updated 8mo ago

Evaluation results

MethodLinks
Proposed modelSize (after training)=44.85MB, Trainable parameters=3.7674M, MAdd=371.93M2020.0999.05
MobileNet (v2)Size (after training)=49MB, Trainable parameters=4.2M, MAdd=575M2020.0998.89
NASNetSize (after training)=64MB, Trainable parameters=4.2M, MAdd=567M2020.0998.73
ShuffleNet (v2)Size (after training)=46.9MB, Trainable parameters=5.4M, MAdd=564M2020.0998.27
AlexNetSize (after training)=294MB, Trainable parameters=25.704M, MAdd=725M2020.0998.237
VGG16Size (after training)=1.5GB, Trainable parameters=138M, MAdd=154.7G2020.0997.6