Center for Vision, Speech and Signal Processing

Center for Vision, Speech and Signal Processing
Name	Center for Vision, Speech and Signal Processing
Established	1990s
Type	Research institute
Location	Guildford, Surrey
Parent	University of Surrey
Fields	Computer vision, Speech recognition, Signal processing

Contents

Center for Vision, Speech and Signal Processing is an academic research institute associated with University of Surrey located in Guildford, Surrey, United Kingdom. The centre conducts interdisciplinary work that intersects computer science, electrical engineering, mathematics, psychology, and neuroscience through collaborations with international institutions such as Massachusetts Institute of Technology, Stanford University, and Imperial College London. Faculty and researchers have contributed to initiatives connected with organizations including European Research Council, Engineering and Physical Sciences Research Council, European Union, Defence Science and Technology Laboratory, and NVIDIA.

History

The centre originated during expansion of research at University of Surrey in the late 1990s and early 2000s, drawing expertise from Royal Holloway, University of London, Oxford University, and Cambridge University. Early work built on traditions from Bell Labs, AT&T Laboratories, and SRI International, while interacting with projects at European Space Agency and CERN. Over time the centre engaged with initiatives led by Google Research, Microsoft Research, Facebook AI Research, and IBM Research, and researchers presented at conferences such as CVPR, ICCV, ECCV, ICASSP, NeurIPS, and ACL.

Research spans computer vision topics including object recognition, image segmentation, scene understanding, and 3D reconstruction; signal domains including audio signal processing, speech enhancement, array processing, and sensor fusion; and machine learning themes such as deep learning, probabilistic graphical models, Bayesian inference, and reinforcement learning. Applications link to autonomous vehicles, medical imaging, biometrics, remote sensing, robotics, human–computer interaction, and augmented reality. Methodological collaborations reference techniques from convolutional neural network, transformer architecture, support vector machine, and hidden Markov model research lines developed at University of Toronto, Carnegie Mellon University, ETH Zurich, and École Polytechnique Fédérale de Lausanne.

Laboratories host hardware such as GPU clusters from NVIDIA, FPGA boards, and real-time systems compatible with ROS and PX4. Imaging suites contain high-resolution cameras from Sony, Canon, and depth sensors from Microsoft Kinect and Intel RealSense. Audio labs include arrays and microphones used in studies related to Bose Corporation devices and platforms developed by Apple Inc. and Google. Computational resources integrate cloud credits from Amazon Web Services, Google Cloud Platform, and Microsoft Azure and are used alongside datasets such as ImageNet, COCO (dataset), LibriSpeech, VoxCeleb, and KITTI (dataset).

The centre offers postgraduate research supervision tied to PhD, MPhil, and MSc programs at University of Surrey, and contributes modules affiliated with Faculty of Engineering and Physical Sciences. Training includes doctoral training partnerships with UK Research and Innovation schemes and summer internships in collaboration with companies such as DeepMind, OpenAI, Waymo, and Siemens Healthineers. Workshops and short courses have been delivered at venues like Royal Society, IEEE Signal Processing Society meetings, and European Conference on Machine Learning events, while alumni have taken positions at Amazon Web Services, Tesla, Adobe Systems, Spotify, Palantir Technologies, and Zebra Technologies.

Notable projects include advances in robust speech recognition influential to Amazon Alexa, noise-robust systems used in NATO communications trials, computer vision algorithms applied to medical imaging collaborations with NHS England hospitals, and autonomous perception modules trialed with Transport for London and Highways England. Publications and tools have been cited alongside landmark works from Yann LeCun, Geoffrey Hinton, Yoshua Bengio, Fei-Fei Li, and Andrew Ng and have been demonstrated at venues such as Royal Institution lectures and BBC science features. Outcomes have informed standards and benchmarks maintained by groups like IEEE, ISO, and ETSI and contributed to patents filed with United States Patent and Trademark Office and European Patent Office.