Current Research

CatNet: Class Incremental 3D ConvNets for Lifelong Egocentric Gesture Recognition

We proposed a lifelong learning framework for EgoGesture video recognition.

Paper and Code

Generative Adversarial Networks in Computer Vision: A Survey and Taxonomy

We reviewed current advances of generative adversarial networks (GANs) from aspects of loss function and architectures.

arXiv:1906.01529 [paper]

GitHub page: Code

GAN paper list: paper list

A Brain-inspired Evaluation Metric for Generative Adversarial Networks.

We demonstrated the efficacy of a convolutional neural network based framework trained by using human's neural signal for evaluating generative adversarial networks .

arXiv:1905.04243 [paper]

Use of Neural Signals to Evaluate the Quality of Generative Adversarial Network Performance in Facial Image Generation.

We introduced using human being's neural signals to evaluate generative adversarial networks.

Cognitive Computation [paper]

Quick and Easy Time Series Generation with Established Image-based GANs.

We introduced a quick way by using image-based GANs to generate time series data. ECG and PPG data have been demonstrated using this approach.

arXiv preprint:1902.05624 [paper]

Spatial filtering pipeline evaluation of cortically coupled computer vision system for rapid serial visual presentation.

We evaluated current spatial filtering pipeline for the cortically coupled computer vision system and we proposed a new spatial filtering technique based on LDA beamformer.

Brain-Computer Interfaces. 5:132–145, 2018 [paper]

An Interpretable Machine Vision Approach to Human Activity Recognition using Photoplethysmograph Sensor Data.

We demonstrated using a machine vision approach for classifying the PPG data and showed interpretation of the used deep neural network.

Irish Conference on Artificial Intelligence and Cognitive Science, 2018 [paper]