Summary of Deep Convolutional Neural Networks on Multiclass Classification Of Three-dimensional Brain Images For Parkinson’s Disease Stage Prediction, by Guan-hua Huang et al.
Deep Convolutional Neural Networks on Multiclass Classification of Three-Dimensional Brain Images for Parkinson’s Disease Stage Prediction
by Guan-Hua Huang, Wan-Chen Lai, Tai-Been Chen, Chien-Chin Hsu, Huei-Yung Chen, Yi-Chen Wu, Li-Ren Yeh
First submitted to arxiv on: 31 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This study develops a machine learning model to accurately predict Parkinson’s disease stages using functional medical imaging techniques such as single-photon emission computed tomography (SPECT). The researchers utilize two SPECT datasets from different hospitals, containing 634 and 202 subjects respectively. They experiment with various model architectures, treating the 3D brain images as sequences of 2D slices and feeding them into 2D convolutional neural network (CNN) models pretrained on ImageNet. Additionally, they apply 3D CNN models pretrained on Kinetics-400 and incorporate an attention mechanism to account for varying slice importance. The results show that 2D models pretrained on ImageNet outperform 3D models pretrained on Kinetics-400, while models with the attention mechanism perform better than both 2D and 3D models. Furthermore, cotraining the two datasets using weight sharing proves effective in improving model performance when the datasets are sufficiently large. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study is trying to help doctors diagnose a condition called Parkinson’s disease more accurately. They’re using special pictures of the brain taken from outside the body (called SPECT scans) to train computers to predict how bad the disease is. The researchers tested different ways of using these pictures and found that one way worked better than others. They also tried combining two sets of pictures together, which made their computer model even more accurate. This could help doctors give patients a more precise diagnosis and start treatment sooner. |
Keywords
» Artificial intelligence » Attention » Cnn » Machine learning » Neural network
