Image and vision computing
The theory and fundamental underpinning of Image and Vision Computing in both 2D and 3D, across the electromagnetic spectrum, with applications including: feature and pattern analysis, pattern recognition, computer-based image interpretation (excluding medical image processing - see the Medical Imaging research area), document image processing, video image processing and analysis, image databases, machine vision and robotics.
Our strategy recognises its importance to data science and in underpinning research across sectors such as robotics, security, defence, healthcare, communications, creative industries, media and manufacturing.
By the end of the current Delivery Plan, we aim to have:
- A research area which continues to address growing relevance to data science, for analysing, understanding and exploiting the ever-increasing visual data and information (including multi-modal data) generated by broadcast media, social media and other sources; and a portfolio which further explores and exploits the links between language and vision, addressing challenges in multi-modal interface research
- An increase in collaboration by researchers working in this area with robotics and autonomous systems, in terms of making a strong contribution to scene understanding for robot vision (together with the research areas Vision, Hearing and Other Senses, and Artificial Intelligence Technologies)
- A research area which continues to reflect a high degree of relevance to the creative industries, working with Graphics and Visualisation, in terms of image capture and analysis for more realistic rendering of scenes and faces; this research area will also be maintained to reflect the contribution that Image and Vision Computing will continue to play in developing biometric technologies and its importance in security and defence
- A portfolio which continues to meet the demand for core capability and underpinning Image and Vision Computing research across many sectors: robotics, security, defence, healthcare (e.g. in activity monitoring in the home environment), communications, environment, manufacturing, media and the creative industries
- Researchers who consider what contribution they can make to the Medical Imaging research area, in the context of understanding and extracting clinically relevant information from medical images
- A portfolio with a greater proportion of early-career researchers, to ensure the longer-term health of this research area.
Researchers working in this area should play an important role in delivering the objectives of EPSRC's Future Intelligent Technologies and Data Enabled Decision Making cross-ICT priorities, and are also well-placed to contribute to the other EPSRC cross-ICT priorities. To maximise the impact of these contributions, Image and Vision Computing researchers should ensure effective communication with researchers in other contributing areas (e.g. Artificial Intelligence Technologies, Graphics and Visualisation, Vision, Hearing and Other Senses, and Human-Computer Interaction).
Researchers should acknowledge and demonstrate the importance of responsible innovation in their proposals, in relation to addressing the issues of trust, identity and privacy when using visual data from large-scale social networks and surveillance sources, and where there are issues in using technology which will need to be accepted and trusted by the general public (e.g. computer vision systems in driverless cars)Highlights:
This has been a strong area of research in the UK for many years, with a number of leading groups located here. EPSRC funding, and funding from other sources such as the European Union, is distributed across many UK institutions. Researchers are publishing competitively at leading international conferences.
Because of the UK's strong research position, there are strong industrial links to large corporations such as the BBC, Microsoft, Google, Facebook, Apple and IBM, and to a good number of small and medium-sized enterprises (SMEs). There has been huge growth in industrial research in this area, with Google, Amazon and Facebook heavily recruiting UK researchers. Google acquired the Vision Factory (an Oxford University spinout) in 2014.
This is a significant research area for advances in data science, for organising and understanding large amounts of visual information, and in developing context-based information retrieval of images, video and graphics from sources such as broadcast and social media (working together with the research areas Artificial Intelligence Technologies and, increasingly, Natural Language Processing). It is also significant for advances in robotics and autonomous systems, in terms of scene understanding in robot vision, in many robotics application areas (working together with the research areas Artificial Intelligence Technologies and Vision, Hearing and Other Senses). (Evidence source 3)
Although UK universities have a strong history of supporting research in this area, retention of doctoral graduates in academia is in question; as well as heavy industrial recruitment, they tend to be attracted to universities abroad that are creating new programmes. UK academic institutions risk being depleted of a postdoctoral workforce and therefore of future leadership. (Evidence source 3)
We need to ensure we have appropriate provision to cater for the demand for trained people with skills sufficient for the academic and industrial pipeline. There is a continuing need and demand for underpinning Image and Vision Computing research across many sectors (see list in 'Strategic Focus' above). (Evidence source 4,5)
There is also a need for specialist High Performance Computing for large-scale analysis and processing of large image/video databases. By 2019, annual global Internet Protocol traffic is predicted to reach 2 zettabytes, with 80% of this predicted to be video (up from 64% in 2014). (Evidence source 6) In 2015, video accounted for 55% of total mobile data traffic and this is predicted to rise to 75% in 2020. Visual information is the key driver for both internet and mobile connectivity. (Evidence source 6)
The main research areas that connect with this one are: Artificial Intelligence Technologies; Graphics and Visualisation; Human-Computer Interaction; Medical Imaging (which includes Image and Vision Computing for medical applications); Robotics; Information Systems; and Vision, Hearing and Other Senses. Other connecting research areas include Natural Language Processing and Speech Technology.
This research area contributes to all four Prosperity Outcomes and specifically to the following Ambitions within Connected, Productive and Resilient Nation Outcomes:
C1: Enable a competitive, data-driven economy
This research area is expected to be key to delivering this, by utilising visual data analytics to gain insights into large amounts of visual data and delivering intelligent technologies and systems by developing scene understanding for robot vision.
P4: Drive business innovation through digital transformation
This research area is expected to contribute to the development of advanced digital technologies – specifically robotics, intelligent systems and advanced data analytics.
R3: Develop better solutions to acute threats: cyber, defence, financial and health
This research area is expected to contribute in terms of improved visual data analytics for security and defence.
- Input from the ICT, Engineering and Healthcare Strategic Advisory Teams, the UK Computing Research Committee (UKCRC) Executive Committee and Research Excellence Framework (REF) 2014 panellists.
- Analysis of EPSRC funding data.
- Community engagement (individual input, group feedback and team visits).
- UK Government, Government Department Reports, (2016).
- Annual Report of the Government Chief Scientific Adviser, Forensic Science and Beyond, (2015).
- Cisco, Cisco Visual Networking Index: Forecast and Methodology, (2016).
Research area connections
This diagram shows the top 10 connections between Research Areas within the EPSRC research portfolio. The depth of the segment relates to value of grants and the width of the segment relates to the number of grants shared by those two Research Areas. Please click to see the related Research Area rationale.
Visualising our Portfolio (VoP)
Visualising our portfolio (VoP) is a tool for users to visually interact with the EPSRC portfolio and data relationships.
EPSRC support by research area in Image and vision computing (GoW)
Search EPSRC's research and training grants.