Jun Jet Tai

Jun Jet Tai

About

7
Publications
2,769
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
4
Citations

Publications

Publications (7)
Article
Obstacle avoidance and navigation (OAN) algorithms typically employ offline or online methods. The former is fast but requires knowledge of a global map, while the latter is usually more computationally heavy in explicit solution methods, or is lacking in configurability in the form of artificial intelligence (AI) enabled agents. In order for OAN a...
Chapter
Tai, Jun JetInnocente, Mauro SebastiánMehmood, OwaisIn this work, a novel high-speed railway fastener detector is introduced. This fully convolutional network, dubbed FasteNet, foregoes the notion of bounding boxes and performs detection directly on a predicted saliency map. FasteNet uses transposed convolutions and skips connections, the effective...
Preprint
Full-text available
In this work, a novel high-speed railway fastener detector is introduced. This fully convolutional network, dubbed FasteNet, foregoes the notion of bounding boxes and performs detection directly on a predicted saliency map. Fastenet uses transposed convolutions and skip connections, the effective receptive field of the network is 1.5$\times$ larger...
Conference Paper
Full-text available
Obstacle Avoidance and Navigation (OAN) algorithms are an active research field dominated by either offline or online methods. The former method is fast but requires a prior known map while the latter method can function without a prior known map at the expense of high computational requirements. To bring OAN algorithm to mass produced mobile robot...
Preprint
In this work, a novel high-speed single object tracker that is robust against non-semantic distractor exemplars is introduced; dubbed BOBBY2. It incorporates a novel exemplar buffer module that sparsely caches the target's appearance across time, enabling it to adapt to potential target deformation. As for training, an augmented ImageNet-VID datase...
Article
Full-text available
Inverted pendulum remained as the most popular topic for control theory researches because of its characteristic of being non-linear, unstable and under-actuated system. It is ideal for verification, validation and enhancement of control theory by stabilizing the inverted pendulum in an upright position using various controller and stabilizer mecha...

Network

Cited By

Projects

Project (1)
Project
The aim of the project is two-fold: 1- Develop algorithms for drones to trace railway tracks and roads autonomously. 2- Develop Artificial Intelligence algorithms to equip the drones with cognitive learning capabilities to identify maintenance issues and inform decision-makers (e.g. via computer vision).