# Ischemic Stroke Lesion Segmentation in CT Perfusion Scans using Pyramid Pooling and Focal Loss

arXiv: https://arxiv.org/abs/1811.01085 · HTML: https://mlsanity.com/paper/1811.01085/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 2 Nov 2018.

## Built on by (most cited first, 7 of 7)

- [Multi-input segmentation of damaged brain in acute ischemic stroke patients using slow fusion with skip connection](https://mlsanity.com/paper/2203.10039/) (2022)
- [Transformers-based architectures for stroke segmentation: A review](https://mlsanity.com/paper/2403.18637/) (2024)
- [Deep models for stroke segmentation: do complex architectures always perform better?](https://mlsanity.com/paper/2403.17177/) (2024)
- [CT Perfusion is All We Need: 4D CNN Segmentation of Penumbra and Core in Patients With Suspected Ischemic Stroke](https://mlsanity.com/paper/2303.08757/) (2023)
- [Weakly supervised semantic segmentation of tomographic images in the diagnosis of stroke](https://mlsanity.com/paper/2109.01887/) (2021)
- [Ischemic Stroke Lesion Prediction using imbalanced Temporal Deep Gaussian Process (iTDGP)](https://mlsanity.com/paper/2211.09068/) (2022)
- [Evaluation of Data Augmentation and Loss Functions in Semantic Image Segmentation for Drilling Tool Wear Detection](https://mlsanity.com/paper/2302.05262/) (2023)

## Abstract (arXiv)

We present a fully convolutional neural network for segmenting ischemic stroke lesions in CT perfusion images for the ISLES 2018 challenge. Treatment of stroke is time sensitive and current standards for lesion identification require manual segmentation, a time consuming and challenging process. Automatic segmentation methods present the possibility of accurately identifying lesions and improving treatment planning. Our model is based on the PSPNet, a network architecture that makes use of pyramid pooling to provide global and local contextual information. To learn the varying shapes of the lesions, we train our network using focal loss, a loss function designed for the network to focus on learning the more difficult samples. We compare our model to networks trained using the U-Net and V-Net architectures. Our approach demonstrates effective performance in lesion segmentation and ranked among the top performers at the challenge conclusion.