ProfessorAnil Bharath

I am Professor of Biologically-Inspired Computation & Inference in the Department of Bioengineering at Imperial. I gained my first degree in Electrical & Electronic Engineering from UCL, followed by a PhD from the Biomedical Systems Group in the Department of Electrical and Electronic Engineering at Imperial College. From 2024-2025, I was Founding Academic Director of Imperial Global: Singapore and, with Professor Liu Yang of Nanyang Technological University, I continue to co-lead the IN-CYPHER research programme into the security of data and devices used in delivering AI-driven personalised healthcare.

My research interests have included convolution-based architectures for visual processing since the mid 1990s, and my current research is focussed on the use of deep networks for inference. My research group covers several different areas of machine learning, and has published on reinforcement learning, generative modelling and adversarial learning. A recurring theme of interest to me is the nature of representations learned by deep neural networks; linked to the topic of representation learning, this interest is motivated by the closeness of certain behaviours found in both biological (spiking) and non-biological (artificial) networks of neurons.

I have published in the fields of pattern recognition, machine learning and signal processing, and have extensive experience in the field of traditional "shallow" computer vision. In 1998, I demonstrated a the use of 2-dimensional steerable filters (convolution-based architectures, with feedback) to shape detection, and this has been a useful idea that has been used in different areas of spatial analysis within my group. Later contributions included the use of Bayesian marginalisation in very early-stage computer vision, fast techniques for focus-of-attention, and custom-designed wavelet transforms to analyse images in a scalable manner.

In 2002, (together with Maria Petrou), I initiated the Basic Technology Project "Reverse Engineering Human Visual Processes", which created a blueprint for a scalable subset of processes in the human visual system, particularly of visual area V1. In 2008, myself and Dr Jeffrey Ng spun out the company Cortexica Vision Systems, which applies simplified models of the behaviour of biological visual neurons to the technology of visual search. Cortexica used cloud-based GPUs to offer visual search services to industry and retail, as early as 2010 using autoencoder-based representations designed using "perfect-reconstruction" properties proposed in the field of signal processing. Cortexica was acquired by Zebra Technologies in 2019.

I served as President of the City and Guilds College Association (CGCA - the largest independent Alumni society of the Faculty of Engineering at Imperial College London), from 2022-2024, preceded by Atula Abyesekera, and succeeded by Kelvin Higgins.

My research publications can be found at the tab above, or on Google Scholar.