Skip to main content

Bio-inspired CODECs using steganography

Conference: Verification Futures 2025 (click here to see full programme)
Speaker: Dr Pedro Machado
Presentation Title: Bio-inspired CODECs using steganography
Abstract:

The Bio-inspired CODECs using steganography talk introduces a bio-inspired neural encoder-decoder (CoDec) system that employs Spiking Neural Networks (SNNs) and Steganography to achieve high-efficiency audio compression. By mimicking biological auditory processing, the method encodes sound into sparse spike-based representations and embeds them into images, enabling compact storage and transmission. While the approach demonstrates significant compression gains, its current software implementation suffers from computational bottlenecks in real-time encoding and decoding. To address this limitation, we explore hardware acceleration using FPGAs and neuromorphic processors, which leverage parallel and event-driven architectures to dramatically improve processing speed. The discussion focuses on optimising the CoDec pipeline for embedded and edge applications, paving the way for scalable, energyefficient bio-inspired compression systems.

Speaker Bio:

Dr. Pedro Machado, a Senior Lecturer in Computer Science and Course Leader for the MSc in Artificial Intelligence, pioneers next-generation AI systems that decode underwater ecosystems by merging Spiking Neural Networks (SNNs), neuromorphic computing, and bio-inspired sensing. His work mimics biological neural processes (from retinal ganglion cells optimising contrast detection to macrophages-inspired immune-response patterns) to enable ultra-efficient, real-time analysis of aquatic behavior. By leveraging neuromorphic edge AI, such as Intel Loihi 2 and BrainChip Akida hardware, his systems deploy event-driven vision and sound processing in remote underwater environments, achieving high compression rates in bioacoustic data for sustainable, long-term monitoring. These innovations detect subtle shifts in fish behavior, predicting stress, disease, or environmental threats like complex gill disease before visible symptoms emerge. Bridging computational neuroscience, aquaculture and conservation, Dr. Machado’s research delivers scalable, energy-efficient tools for precision aquaculture, pollution tracking, and habitat restoration, transforming how we safeguard biodiversity and prevent ecosystem collapse in oceans and freshwater systems.

Key Points:
  • Bio-inspired Compression with Steganography
  • High Efficiency, Current Software Bottlenecks
  • Hardware Acceleration for Real-time Performance
  • Close Menu