Video Capsule Endoscopy Classification using Focal Modulation Guided Convolutional Neural Network

2.) Overview

2.1.)Introduction

In our paper, we have proposed FocalConvNet, a focal modulation network integrated with lightweight convolutional layers for the classification of small bowel anatomical landmarks and luminal findings. FocalConvNet leverages focal modulation to attain global context and allows global-local spatial interactions throughout the forward pass. Moreover, the convolutional block with its intrinsic inductive/learning bias and capacity to extract hierarchical features allows our FocalConvNet to achieve favourable results with high throughput.

2.2.) Model Architecture of Our FocalConvNet

3.) Training and Testing

3.1)Data Preparation

Follow the data preparation procedure in the official dataset repository "Kvasir-Capusle"

3.2)Training and testing

1.) The architecture for the FocalConvNet is defined in focalconv.py 2.) run the training script in the official dataset repo and replace the model definition in Line 325 with the architecture of FocalConvNet.

4.) FAQ

Please feel free to contact me if you need any advice or guidance in using this work (E-mail)

Acknowlegment

For our codebase, we use the repo of Focal Modulation networks and Kvasir-Capusle. We thank the authors for the nicely organized code!