SpFormer: Spatio-Temporal Modeling for Scanpaths with Transformer

Authors: Wenqi Zhong, Linzhi Yu, Chen Xia, Junwei Han, Dingwen Zhang

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct extensive experiments on four databases under three tasks.
Researcher Affiliation Academia School of Automation, Northwestern Polytechnical University, China wenqizhong@mail.nwpu.edu.cn, 15160557827@mail.nwpu.edu.cn, cxia@nwpu.edu.cn, junweihan2010@gmail.com, zhangdingwen2006yyy@gmail.com
Pseudocode No The paper describes methods using text and mathematical equations, but does not include any explicitly labeled 'Pseudocode' or 'Algorithm' blocks.
Open Source Code Yes The code can be obtained from https://github.com/wenqizhong/Sp Former.
Open Datasets Yes We apply two datasets, Saliency4ASD (Duan et al. 2019) dataset and our collected dataset, to evaluate the ASD recognition performance of the Sp Former. ... We utilize a toddler age prediction (TAP) dataset1 obtained from (Dalrymple et al. 2019)... 1https://osf.io/ugvj4 ... Koehler et al. (2014) proposed a visual perceptual task (VPT) dataset2... 2https://data.mendeley.com/datasets/8rj98pp6km/1
Dataset Splits No The paper states 'Utrain denotes the subject set for training' and refers to previous experimental protocols for settings, but does not provide specific numerical train/validation/test dataset splits within the provided text.
Hardware Specification No The paper does not provide any specific hardware specifications such as GPU models, CPU types, or memory used for running the experiments.
Software Dependencies No The paper does not provide specific software dependencies or their version numbers (e.g., Python, PyTorch, or CUDA versions) required for replication.
Experiment Setup Yes We set α = σ = 0.5, β = 1 σ = 0.5 in Eq. (20), and λ = 0.1 in Eq.( 24).