Boxed up and good to go. This blog post is meant for annotators who are embarking on their journey to help label data for object detection. We hope to briefly introduce you to the computer vision task and your task as an annotator. 1. Introduction Object detection is a computer vision task…

Need to build a web prototype but no idea where to start? No worries! Ernest is here to the rescue. This post will provide you with some starter code using SOTA popular frameworks which you should have no problem finding stackoverflow answers to your questions. Disclaimer I am no expert, but…

Object detection models on videos underperform when challenged by motion blur and occlusions. We investigated the use of temporal information to boost detection performance. This was a project by Amanda Koh for her internship at Digital Hub, supervised by Evan Ling, Lee Ying Hui, Teo Jin Ming. Background Currently, when performing…

We explore the use of labelled data to construct a model that can be used to flag anomalous, future unlabeled data, in the context of flight trajectories. A project by Ryan Tan during his internship at Digital Hub. With so much data out there, we are finding an increasing number…

We explore different machine learning techniques for achieving the goal of pathological voice classification and proposes the use of a deep-learning model called RawNetThis was a project by Gabriel Ng during his internship at Digital Hub; supervised by Jax. Since the covid-19 pandemic began, the significance of pathological voice classification…

basic overview to working well on git, in order to maintain useful versioning. The general idea is to branch when adding features or debugging, to submit a pull request when done. This ideally enables us to maintain a working version of our project on main at all times. Finding your…

Hypebeast 21’st century fashion can be confusing. Teenagers queueing for hours outside branded boutiques to buy thousand-dollar kicks (shoes), only to pose for some photos in them before selling them on the internet for a surprisingly handsome profit. Clothes, bags, caps, shovels, and even kayaks spike in value when they are…

A primer on adversarial attacks and synthetic image generation prior to the lowdown on robustness testing of vessel classifiers to natural perturbations like weather and changes in camera angle, with the use of synthetic image generation using 3DB. This was a project by Wang Li Rong during her internship at…

You take the blue pill, the story ends, you wake up in your bed and believe whatever you want to believe. You take the red pill, you stay in wonderland, and I show you how deep the rabbit hole goes. -Morpheus Whether you’re an intern trying to run ROS without…