Beta's projects for MSc students 2013—2014

General

Speaker Recognizer

1 student, maximum 3 groups | Development project | Updated 2013-11-01[5]

Description

Build a system that recognizes and labels the speakers in a recorded radio talk show or phone-in show.

Preferably, the system should be language-indpendent.

A more advanced version of the system should be able to run in stream mode and takes live speech, generating output or as the input is analyzed, and possibly correcting earlier outputs when necessary.

Note that the student is expected to build their own collection of training and testing data.

The system is expected to be implemented in Linux or OS X.

Requirements

Deliverables

Bird Extractor

1 student, maximum 3 groups | Development project | Updated 2013-11-01[5]

Description

Build and train a system that when given a photo of bird(s), it can identify the locations of its body parts such as eyes, bill, tarsus, and wing feathers of different types (primary, secondary, coverts) whenever possible.

Ideally, the outlines of the birds in the photo should be identified.

Note that the student is expected to build their own collection of training and testing data.

The system is expected to be implemented in Linux or OS X.

Requirements

Deliverables

Financial data forecaster

1 student, maximum 5 groups | Research project | Updated 2013-11-01[5]

Description

Given the historical data of a number of stock or index prices at different points in time, design an algorithm that would predict their values in the future.

Some factors the algorithm can take into the consideration include the day in month, weekday of day, time of day, various financial indicators, correlations between data from different time series.

Time series data of more than 30 stocks and/or indices are expected to be analyzed.

Note that the student is expected to build their own collection of training and testing data.

Requirements

Deliverables

Automatic classification of radar-observed precipitation systems

1 student, maximum 2 groups | Research+Development project | Under the "Cooperative training programmes" between HKU CS and HKO | Updated 2013-11-01[5]

Description

Thunderstorms are known to display different levels of spatial organisation (isolated cells, clusters, lines, etc) under specific meteorological conditions, each capable of bringing particular weather hazards. In this project, the student will develop algorithms for automatic classification of such precipitation systems as observed on a weather radar. Conditions favourable for formation of different types of thunderstorms along the South China coastal areas will then be explored by analysis of 3-dimensional meteorological data fields.

Requirements

Deliverables

Automatic 4-dimensional identification of weather features

2 to 3 students, maximum 2 groups | Research+Development project | Under the "Cooperative training programmes" between HKU CS and HKO | Updated 2013-11-01[5]

Description

Despite rapid and continual development in Numerical Weather Prediction (NWP) techniques, subjective assessment of weather patterns based on analysis of weather systems by human forecasters still plays a key role in modern operational forecasting. In this project, the student will develop automatic algorithms for identifying and tracking key surface and upper-air weather systems from 4-D meteorological data fields spanning East Asia and the Western North Pacific, both actual and forecast, with a view to aiding real-time weather analysis by forecasters.

Requirements

Deliverables

References