EAR-UNet: A deep learning-based approach for segmentation of tympanic membranes from otoscopic images

Keyphrases

• Deep Learning Methods 100%
• Tympanic Membrane 100%
• U-Net 100%
• Encoder 50%
• Decoder 50%
• New Loss Function 50%
• EfficientNet 50%
• Skip Connection 50%
• Attention Gate 50%
• ResNet Architecture 50%
• Residual Block 50%
• Connection Paths 50%
• Image-based 25%
• Neural Network 25%
• State-of-the-art Techniques 25%
• Otitis Media 25%
• Automatic Segmentation 25%
• EfficientNet-B4 25%
• Overfitting 25%
• Segmentation Performance 25%
• U-Net Architecture 25%
• Dice Similarity Coefficient 25%
• Stochastic Weight Averaging 25%
• Network Training 25%
• Post-processing Techniques 25%
• Fully Convolutional Neural Network 25%
• Segmentation Problem 25%
• Local Minima 25%
• Shape Distance 25%
• Reception Field 25%

Engineering

• Attention Gate 100%
• Loss Function 100%
• Deep Learning 100%
• Experimental Result 50%
• Tasks 50%
• Processing Step 50%
• Similarities 50%
• Ground Truth 50%
• Convolutional Neural Network 50%
• Local Minimum 50%
• Network Training 50%
• State-of-the-Art Method 50%

Computer Science

• Deep Learning 100%
• Residual Neural Network 100%
• Experimental Result 50%
• Neural Network 50%
• Preprocessing 50%
• Convolutional Neural Network 50%
• Segmentation Task 50%
• Segmentation Performance 50%
• Automatic Segmentation 50%
• Postprocessing Step 50%
• Local Minimum 50%