Abstract:In modern cyber-physical systems, the integration of AI into vision pipelines is now a standard practice for applications ranging from autonomous vehicles to mobile devices. Traditional AI integration often relies on cloud-based processing, which faces challenges such as data access bottlenecks, increased latency, and high power consumption. This article reviews embedded AI vision systems, examining the diverse landscape of near-sensor and in-sensor processing architectures that incorporate convolutional neural networks. We begin with a comprehensive analysis of the critical characteristics and metrics that define the performance of AI-integrated vision systems. These include sensor resolution, frame rate, data bandwidth, computational throughput, latency, power efficiency, and overall system scalability. Understanding these metrics provides a foundation for evaluating how different embedded processing architectures impact the entire vision pipeline, from image capture to AI inference. Our analysis delves into near-sensor systems that leverage dedicated hardware accelerators and commercially available components to efficiently process data close to their source, minimizing data transfer overhead and latency. These systems offer a balance between flexibility and performance, allowing for real-time processing in constrained environments. In addition, we explore in-sensor processing solutions that integrate computational capabilities directly into the sensor. This approach addresses the rigorous demand constraints of embedded applications by significantly reducing data movement and power consumption while also enabling in-sensor feature extraction, pre-processing, and CNN inference. By comparing these approaches, we identify trade-offs related to flexibility, power consumption, and computational performance. Ultimately, this article provides insights into the evolving landscape of embedded AI vision systems and suggests new research directions for the development of next-generation machine vision systems.

Review of bio-inspired image sensors for efficient machine vision

Recent advances in in-sensor computational vision sensors: from mechanisms to applications

Recent Advances in Bio-Inspired Vision Sensor: A Review

Progress in Bioinspired Photodetectors Design for Visual Information Processing

Bioinspired sensing-memory-computing integrated vision systems: biomimetic mechanisms, design principles, and applications

Survey on Energy-Efficient Deep Neural Networks for Computer Vision

In-sensor computing for machine vision

Biomimetic Optoelectronics with Nanomaterials for Artificial Vision

IEEE Access Special Section Editorial: Biologically Inspired Image Processing Challenges and Future Directions

Bio-inspired Visual Systems Based on Curved Image Sensors and Synaptic Devices

Artificial intelligence-based spatio-temporal vision sensors: applications and prospects

Event-Based Neuromorphic Vision for Autonomous Driving: A Paradigm Shift for Bio-Inspired Visual Sensing and Perception.

Recent advances in bioinspired vision sensor arrays based on advanced optoelectronic materials

Intelligent Optoelectronic Devices for Next‐Generation Artificial Machine Vision

Photoelectric Memristor-Based Machine Vision for Artificial Intelligence Applications

Bioinspired in-sensor visual adaptation for accurate perception

Inspiration from Visual Ecology for Advancing Multifunctional Robotic Vision Systems: Bio‐inspired Electronic Eyes and Neuromorphic Image Sensors

From Near-Sensor to In-Sensor: A State-of-the-Art Review of Embedded AI Vision Systems

A Biologically Inspired Neurocomputing Circuit for Image Representation

Next‐generation machine vision systems incorporating two‐dimensional materials: Progress and perspectives

Bioinspired and Low-Power 2D Machine Vision with Adaptive Machine Learning and Forgetting