#Multi-scale-features

IBM

CrossViT

Official implementation of CrossViT. https://arxiv.org/abs/2103.14899

amrzhd

EEG-MSCNN

This project explores the impact of Multi-Scale CNNs on the classification of EEG signals in Brain-Computer Interface (BCI) systems. By comparing the performance of two models, EEGNet and MSTANN, the study demonstrates how richer temporal feature extractions can enhance CNN models in classifying EEG signals

anthonyweidai

SvANet

[TNNLS 2026] SvANet: Exploiting Scale-Variant Attention for Segmenting Small Medical Objects

rentainhe

TRAR-VQA

[ICCV 2021] Official implementation of the paper "TRAR: Routing the Attention Spans in Transformers for Visual Question Answering"