Experiment Configuration

Welcome to the rECGnition Web Server! This form allows you to customize and fine-tune various aspects of rECGnition_v1 model, from dataset selection to architecture design and hyperparameter optimization. Configure your input settings, dynamically add layers to the model, and upload your own data or select from pre-defined datasets. Tailored for researchers and clinicians, this tool provides the flexibility to experiment and iterate seamlessly without the prior knowledge of coding. Fill out the sections below to build and save your model configuration.

1. Your Details

Please enter your full name.

Please enter a valid email address to receive results.

Please enter your affiliation.

2. Dataset

Select the dataset you want to use for training your model, if you want to test it on custom dataset then select upload custom dataset (Note: uploaded dataset should be same as the format defined by physionet i.e similar to MIT-BIH Arrhythmia Dataset).

3. Model Architecture

ECG Processor

Please enter image size for heart beat images.

Please select the pre-trained CNN backbone to be used.

Patient Characteristic Processor

Please enter the number of patient characteristic features.

Dense Layers

Please enter the number of dense layer to be used for Patient Meta data processor.

Late Fusion of ECG and Patient Characterstic Features

Dense Layers

Please add number of dense layers to be used for meta classifier.

Please enter number of arrhythmia classes to be used for final classification.

Please enter the comma seperate arrhythmia annotations used in dataset.

4. Hyperparameters

Please enter valid learning rate.

Please enter batch size.

Please enter number of epoches.

Please enter a valid value for dropout.

Please select optimizer for training your model.