SwInception -- Local Attention Meets Convolutions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hagerman, David, Naeem, Roman, Lindqvist, Jakob, Lindström, Carl, Kahl, Fredrik, Svensson, Lennart
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913170586599424
author Hagerman, David
Naeem, Roman
Lindqvist, Jakob
Lindström, Carl
Kahl, Fredrik
Svensson, Lennart
author_facet Hagerman, David
Naeem, Roman
Lindqvist, Jakob
Lindström, Carl
Kahl, Fredrik
Svensson, Lennart
contents Sparse vision transformers have gained popularity as efficient encoders for medical volumetric segmentation, with Swin emerging as a prominent choice. Swin uses local attention to reduce complexity and yields excellent performance for many tasks but still tends to overfit on small datasets. To mitigate this weakness, we propose a novel architecture that further enhances Swin's inductive bias by introducing Inception blocks in the feed-forward layers. The introduction of these multi-branch convolutions enables more direct reasoning over local, multi-scale features within the transformer block. We have also modified the decoder layers in order to capture finer details using fewer parameters. We demonstrate a performance improvement on eleven different medical datasets through extensive experimentation. We specifically showcase advancements over the previous state-of-the-art backbones on benchmark challenges like the Medical Segmentation Decathlon and Beyond the Cranial Vault. By showing that the existing inductive bias in Swin can be further improved, our work presents a promising avenue for enhancing the capabilities of sparse vision transformers for both medical and natural image segmentation tasks. Code and pre-trained weights can be accessed at https://github.com/Eiphodos/SwInception.
format Preprint
id arxiv_https___arxiv_org_abs_2605_29954
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SwInception -- Local Attention Meets Convolutions
Hagerman, David
Naeem, Roman
Lindqvist, Jakob
Lindström, Carl
Kahl, Fredrik
Svensson, Lennart
Computer Vision and Pattern Recognition
Sparse vision transformers have gained popularity as efficient encoders for medical volumetric segmentation, with Swin emerging as a prominent choice. Swin uses local attention to reduce complexity and yields excellent performance for many tasks but still tends to overfit on small datasets. To mitigate this weakness, we propose a novel architecture that further enhances Swin's inductive bias by introducing Inception blocks in the feed-forward layers. The introduction of these multi-branch convolutions enables more direct reasoning over local, multi-scale features within the transformer block. We have also modified the decoder layers in order to capture finer details using fewer parameters. We demonstrate a performance improvement on eleven different medical datasets through extensive experimentation. We specifically showcase advancements over the previous state-of-the-art backbones on benchmark challenges like the Medical Segmentation Decathlon and Beyond the Cranial Vault. By showing that the existing inductive bias in Swin can be further improved, our work presents a promising avenue for enhancing the capabilities of sparse vision transformers for both medical and natural image segmentation tasks. Code and pre-trained weights can be accessed at https://github.com/Eiphodos/SwInception.
title SwInception -- Local Attention Meets Convolutions
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.29954