This code implements the image-viewing based ASD screening model proposed in the paper "Attention-based Autism Spectrum Disorder Screening with Privileged Modality". It is used to reproduce the results on the Saliency4ASD dataset. The high-level architecture of the proposed model is visualized below:

If you use our code or data, please cite our paper:

@InProceedings{Chen_2019_ICCV,
author = {Chen, Shi and Zhao, Qi},
title = {Attention-Based Autism Spectrum Disorder Screening With Privileged Modality},
booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
month = {October},
year = {2019}
}

1. Requirements for Pytorch. We use Pytorch 1.2.0 in our experiments.
2. Requirements for Tensorflow. We only use the tensorboard for visualization.
3. Python 3.6+

The code performs leave-one-subject-out evaluation on the training splits of Saliency4ASD dataset. Please download the dataset accordingly and unzip it to folder saliency4asd.

Running the experiments with our code is straightforward, as the default parameters have already been set following the paper, simply call:

python main.py --checkpoint_path $CHECKPOINT_DIR

The tensorboard visualization (stored in $CHECKPOINT_DIR) provides the prediction accuracy (predicted confidence on the correct labels) on different hold-out subjects during the leave-one-subject-out evaluation.