iSegFormer: Interactive Segmentation via Transformers with Application to 3D Knee MR Images
This work addresses interactive segmentation for 3D knee MR images, representing an incremental improvement in efficiency for a domain-specific application.
The authors tackled interactive 3D medical image segmentation by proposing iSegFormer, a memory-efficient transformer combining a Swin transformer with an MLP decoder, which achieved high computational efficiencies.
We propose iSegFormer, a memory-efficient transformer that combines a Swin transformer with a lightweight multilayer perceptron (MLP) decoder. With the efficient Swin transformer blocks for hierarchical self-attention and the simple MLP decoder for aggregating both local and global attention, iSegFormer learns powerful representations while achieving high computational efficiencies. Specifically, we apply iSegFormer to interactive 3D medical image segmentation.