Automatic description generation from digital video

Video captioning visualization result on VAL dataset [Image by Author].

GitHub repository: https://github.com/bozliu/video-captioning

Abstract

Recently, interactive media material consisting of the medium text, audio, picture and video intrinsically has a multimedia style. With the assistance of advanced deep learning (DL) methods, some of the exceptional computer vision (CV) issues have been successfully resolved. Video Captioning is automatic description generation from digital video. In this article, a novel deep learning neural network architecture is introduced and implemented on the MST-VTT dataset by using I3D and 2DCNN to extract the features and fuse them with spatial, temporal and channel attention mechanism. …


To introduce some novel data augmentation methods on improving the resolution of a single image

1 Challenge Description

Figure 1. An example of single image super-resolution [Image by author].

The goal of this mini challenge is to increase the resolution of a single image (by four times). The data for this task comes from the DIV2K dataset [1]. For this challenge, we prepared a mini-dataset, which consists of 500 training and 80 validation pairs of images, where the HR images have 2K resolution and the LR images are downsampled four times.

For each LR image, algorithms will increase the resolution of the images. The quality of the output will be evaluated based on the PSNR between the output and HR images. …


How to deal with domain gap and imbalanced data in a large dataset.

GitHub Repository: https://github.com/bozliu/CelebA-Challenge

1 Challenge Description

The goal of this challenge is to identify 40 different attribute labels depicted in a facial photograph. The data for this task comes from the CelebA dataset [1] and face images has been pre-cropped and aligned to make the data more manageable.

Figure 1. An example of part of attribute labels on a human face picture [Image by Author].

2 The Problem of Domain Gap and Imbalanced Data


Gender and age classification for single images

Team Member: LIU BOZHONG, ZHOU JIANAN, ZHU ZHICHENG

Time: November 2020

Github Repository: https://github.com/bozliu/Adience

Abstract

This project focuses on the gender and age classification for a single image. We use prevalent CNN architectures (ResNet/VGG) as our backbone and try to improve the accuracy. The first main area of experimentation is to try various network structures, to investigate the effect of hyper-parameters and to design options on the performance. The next facet of the project attempts to propose some novel ideas and tries to improve the accuracy. We propose a new loss function called Order Loss, working with cross-entropy loss to obtain…


To develop the algorithm that classifies drugs based on their biological activity

Team Member: LIU BOZHONG, ZHOU JIANAN, ZHU ZHICHENG

Time: November 2020

Homepage of Competition: https://www.kaggle.com/c/lish-moa/overview

GitHub repository : https://github.com/bozliu/Mechanisms-of-Action-Prediction

1 Introduction

1.1 Problem Statement

This is a multi-label classification problem to determine Mechanism of Action (MoA) of a drug. A new technology can measure simultaneously human cells’ responses to drugs in a pool of 100 different cell types within the same samples and therefore ex-ante, cell types that are better suited for a given drug can be identified. The dataset combines gene expression and cell viability data split into testing and training subsets.

The objective is to use the training dataset to develop an algorithm…


To design and develop the best model for daily historical Apple stock prices (open, high, low, close and adjusted prices)

Supervisor: Professor Pan Guangming, Professor at Nanyang Technological Univeristy

Time: May, 2020

Github Repositoty: https://github.com/bozliu/Financial-Data-Forecasting

1 Objective

The objective of this article is to design and develop the best model for the financial data, the daily historical Apple stock prices (open, high, low, close and adjusted prices).

2 Plot Original Time Series Data

The adjusted prices for the daily prices of the Apple stock from February 1, 2002 to January 31, 2017 can be plotted, as shown in Figure 1. It can be observed that the mean value is not zero and the variance is very high. This indicates that the time series is non-stationary with varying mean…


To build a relative accurate model to predict the drug sales time series data

Supervisor: Professor Pan Guangming, Professor at Nanyang Technological Univeristy

Time: March, 2020

Github Repository: https://github.com/bozliu/Seasonal-Data-Forecasting

1 Objective

The objective of this article is to build a time series analysis model to predict the drug sales time series data.

2 Plot Original Time Series Data

The original data for monthly anti-diabetic drug sales in Australia from 1992 to 2008 can be plotted, as shown in Figure 1. There are a clear increasing trend and a strong seasonal pattern that increases in size as the level of the series increases.

Figure 1. Original time series monthly anti-diabetic drug sales data.

3 Box-Cox Transformation

It can also be seen from Figure 1 that there is a small increase in the variance with the level…


How to use machine learning to auto-tune cavity filters.

Time: November 2019 to April 2020

Github Repository: https://github.com/bozliu/Cavity-Optimisation

Automatic fine tuning of cavity filters

Abstract

Cavity filters are a necessary component in base stations used for telecommunication. Without these filters it would not be possible for base stations to send and receive signals at the same time. Today these cavity filters require fine tuning by humans before they can be deployed. This article has designed and implemented a supervised neural network that can predict radius and height of a cylindrical cavity resonator based on resonant frequencies, electric field and magnetic field. Different machine learning methods have been evaluated, such as decision tree, random forest, k-nearest neighbours…


The project of my bachelor thesis achieved during the internship in Bosch China

Supervisor: Dr. Xie Hui Senior Research Scientist of Bosch China

Time: June to September in 2018

Abstract

With the continuing high demand for road transport of goods, accidents can occur during transportation from one place to another, often resulting in the loss of goods. Therefore, it is necessary to monitor the state of goods in real-time, before and after transportation. A real-time volume detection system can be designed and implemented to detect the volume of goods inside a cabinet during transportation, using several remote agents. Each agent is an embedded device with integrated sensors, Arduino microcontroller and GSM enable modules…


A practical solution for indoor navigation.

Supervisor: Dr. Weiping Ni, Lecturer from the Univerity of Nottingham

Time: May to June 2018

I did this project from May to June with Dr. Weiping Ni, who is also my personal tutor (graduated from Cornell University in Electronics and Computer Engineering with her master’s and PhD degree).

Figure 1. One of the tested paths recorded by the device.

1 Overview

During the summer, we focused on motion recognition and position tracking by using an inertial measurement unit (IMU) sensor and an Arduino board. In the process of data acquisition, since we are not dealing with arm movement or fine finger movement for gesture recognitions, we can take full advantage of a relatively…

Bozhong Liu

Master of Science in Artificial Intelligence at Nanyang Technological University | LinkedIn: linkedin.com/in/bozliu | GitHub: https://github.com/bozliu

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store