Detecting Legends in Graphics: A Case study with Public Space Floor Plans

Monash Summer Vacation Scholarship: AUD 3000 (10 weeks from Nov 30th 2020)

​

Link to Apply: Monash Summer Vacation Projects

​

Project Details:

Information graphics can vary from charts, molecule diagrams, house plans to public space floor plans. Most of such graphics contain a legend/key to support interpretation. Graphic recognition researchers have focused on automatically recognising such graphics for varying purposes and one such important purpose is accessible graphic generation. Graphic recognition enables to generate accessible graphics speedily, for a low cost and without requiring manual transcribers. 

Out of diverse information graphics, public space floor plans has received a considerable less attention from accessible researchers. Public spaces such as libraries, museums and train-stations are an integrated part of our daily life. These public spaces can be of different types of buildings such as, 

• General Use – Shopping centres, Museums, Libraries, Hospitals, Banks, Theatres, Hotels

• Public transport – Train stations, Air ports, Passenger Ships

• Public services and municipality buildings - Court buildings, Police stations, Parliament

• Education and Work - University and School buildings, Office buildings

Generating accessible versions of such public space plans using graphic recognition would be complex due to varying symbols. A major step that can considerably reduce this complexity is to recognise their legends. However, legend recognition of information graphics has not received much attention and hence the focus of this project is to address this shortcoming.

The final products of the project would involve a literature review on detecting legends in graphics, a corpus of public space floor plans, a tool to detect legends in public space floor plans and a publication in a recognised venue. 

​

Supervised by: Dr. Anuradha Madugalla

​

Required:

Excellent image processing knowledge

Prior experience in working with image processing libraries - OpenCV or MatLab
Solid programming skills: Python or C++
Excellent communication and writing skills

Currently living in Australia (International students/PR/AU Citizen) 

​