Vision Algorithms for Embedded Vision
Most computer vision algorithms were developed on general-purpose computer systems with software written in a high-level language
Most computer vision algorithms were developed on general-purpose computer systems with software written in a high-level language. Some of the pixel-processing operations (ex: spatial filtering) have changed very little in the decades since they were first implemented on mainframes. With today’s broader embedded vision implementations, existing high-level algorithms may not fit within the system constraints, requiring new innovation to achieve the desired results.
Some of this innovation may involve replacing a general-purpose algorithm with a hardware-optimized equivalent. With such a broad range of processors for embedded vision, algorithm analysis will likely focus on ways to maximize pixel-level processing within system constraints.
This section refers to both general-purpose operations (ex: edge detection) and hardware-optimized versions (ex: parallel adaptive filtering in an FPGA). Many sources exist for general-purpose algorithms. The Embedded Vision Alliance is one of the best industry resources for learning about algorithms that map to specific hardware, since Alliance Members will share this information directly with the vision community.
General-purpose computer vision algorithms
One of the most-popular sources of computer vision algorithms is the OpenCV Library. OpenCV is open-source and currently written in C, with a C++ version under development. For more information, see the Alliance’s interview with OpenCV Foundation President and CEO Gary Bradski, along with other OpenCV-related materials on the Alliance website.
Hardware-optimized computer vision algorithms
Several programmable device vendors have created optimized versions of off-the-shelf computer vision libraries. NVIDIA works closely with the OpenCV community, for example, and has created algorithms that are accelerated by GPGPUs. MathWorks provides MATLAB functions/objects and Simulink blocks for many computer vision algorithms within its Vision System Toolbox, while also allowing vendors to create their own libraries of functions that are optimized for a specific programmable architecture. National Instruments offers its LabView Vision module library. And Xilinx is another example of a vendor with an optimized computer vision library that it provides to customers as Plug and Play IP cores for creating hardware-accelerated vision algorithms in an FPGA.
Other vision libraries
- Halcon
- Matrox Imaging Library (MIL)
- Cognex VisionPro
- VXL
- CImg
- Filters
“LLMs and VLMs for Regulatory Compliance, Quality Control and Safety Applications,” a Presentation from Camio
Lazar Trifunovic, Solutions Architect at Camio, presents the “LLMs and VLMs for Regulatory Compliance, Quality Control and Safety Applications” tutorial at the May 2025 Embedded Vision Summit. By using vision-language models (VLMs) or combining large language models (LLMs) with conventional computer vision models, we can create vision systems that are… “LLMs and VLMs for Regulatory
NVIDIA Opens Portals to World of Robotics With New Omniverse Libraries, Cosmos Physical AI Models and AI Computing Infrastructure
New NVIDIA Omniverse NuRec 3D Gaussian Splatting Libraries Enable Large-Scale World Reconstruction New NVIDIA Cosmos Models Enable World Generation and Spatial Reasoning New NVIDIA RTX PRO Blackwell Servers and NVIDIA DGX Cloud Let Developers Run the Most Demanding Simulations Anywhere Physical AI Leaders Amazon Devices & Services, Boston Dynamics, Figure AI and Hexagon Embrace Simulation and Synthetic Data Generation August 11, 2025—SIGGRAPH—NVIDIA
“Simplifying Portable Computer Vision with OpenVX 2.0,” a Presentation from AMD
Kiriti Nagesh Gowda, Staff Engineer at AMD, presents the “Simplifying Portable Computer Vision with OpenVX 2.0” tutorial at the May 2025 Embedded Vision Summit. The Khronos OpenVX API offers a set of optimized primitives for low-level image processing, computer vision and neural network operators. It provides a simple method for… “Simplifying Portable Computer Vision with
“Quantization Techniques for Efficient Deployment of Large Language Models: A Comprehensive Review,” a Presentation from AMD
Dwith Chenna, MTS Product Engineer for AI Inference at AMD, presents the “Quantization Techniques for Efficient Deployment of Large Language Models: A Comprehensive Review” tutorial at the May 2025 Embedded Vision Summit. The deployment of large language models (LLMs) in resource-constrained environments is challenging due to the significant computational and… “Quantization Techniques for Efficient Deployment
Learn to Optimize Stable Diffusion on Qualcomm Cloud AI 100
This blog post was originally published at Qualcomm’s website. It is reprinted here with the permission of Qualcomm. Dive in to learn how we achieve a 1.4x latency decrease on Qualcomm Cloud AI 100 Ultra accelerators by applying an innovative DeepCache technique to text-to-image generation. What’s more, the throughput can be further improved by 3x
Texas Instruments Demonstration of Edge AI Inference and Video Streaming Over Wi-Fi
The demonstration shows how to use Texas Instruments’ AM6xA to capture live video, perform machine learning, and stream video over Wi-Fi. The video is encoded with H.264/H.265, and streamed via UDP over Wi-Fi using the CC33xx. At the receiver side, the video is decoded and displayed on a screen. The receiver side could be a
“Introduction to Data Types for AI: Trade-offs and Trends,” a Presentation from Synopsys
Joep Boonstra, Synopsys Scientist at Synopsys, presents the “Introduction to Data Types for AI: Trade-offs and Trends” tutorial at the May 2025 Embedded Vision Summit. The increasing complexity of AI models has led to a growing need for efficient data storage and processing. One critical way to gain efficiency is… “Introduction to Data Types for
Machine Vision Defect Detection: Edge AI Processing with Texas Instruments AM6xA Arm-based Processors
Texas Instruments’ portfolio of AM6xA Arm-based processors are designed to advance intelligence at the edge using high resolution camera support, an integrated image sensor processor and deep learning accelerator. This video demonstrates using AM62A to run a vision-based artificial intelligence model for defect detection for manufacturing applications. Watch the model test the produced units as
“Introduction to Radar and Its Use for Machine Perception,” a Presentation from Cadence
Amol Borkar, Product Marketing Director, and Vencatesh Subramanian, Design Engineering Architect, both of Cadence, co-present the “Introduction to Radar and Its Use for Machine Perception” tutorial at the May 2025 Embedded Vision Summit. Radar is a proven technology with a long history in various market segments and continues to plays an increasingly important role in
Optimizing LLMs for Performance and Accuracy with Post-training Quantization
This article was originally published at NVIDIA’s website. It is reprinted here with the permission of NVIDIA. Quantization is a core tool for developers aiming to improve inference performance with minimal overhead. It delivers significant gains in latency, throughput, and memory efficiency by reducing model precision in a controlled way—without requiring retraining. Today, most models
Alif Semiconductor Demonstration of Face Detection and Driver Monitoring On a Battery, at the Edge
Alexandra Kazerounian, Senior Product Marketing Manager at Alif Semiconductor, demonstrates the company’s latest edge AI and vision technologies and products at the 2025 Embedded Vision Summit. Specifically, Kazerounian demonstrates how AI/ML workloads can run directly on her company’s ultra-low-power Ensemble and Balletto 32-bit microcontrollers. Watch as the AI/ML AppKit runs real-time face detection using an
Inuitive Demonstration of On-camera SLAM, Depth and AI Using a NU4X00-based Sensor Module
Shay Harel, Field Application Engineer at Inuitive, demonstrates the company’s latest edge AI and vision technologies and products at the 2025 Embedded Vision Summit. Specifically, Harel demonstrates one of several examples his company presented at the Summit, highlighting the capabilities of its latest vision-on-chip technology. In this demo, the NU4X00 processor performs depth sensing, object
Nota AI Demonstration of Nota Vision Agent, Next-generation Video Monitoring at the Edge
Tae-Ho Kim, CTO and Co-founder of Nota AI, demonstrates the company’s latest edge AI and vision technologies and products at the 2025 Embedded Vision Summit. Specifically, Kim demonstrates Nota Vision Agent—a next-generation video monitoring solution powered by Vision Language Models (VLMs). The solution delivers real-time analytics and intelligent alerts across critical domains including industrial safety,
Nota AI Demonstration of NetsPresso Optimization Studio, Streamlined with Visual Insights
Tairen Piao, Research Engineer at Nota AI, demonstrates the company’s latest edge AI and vision technologies and products at the 2025 Embedded Vision Summit. Specifically, Piao demonstrates NetsPresso Optimization Studio, the latest enhancement to Nota AI’s model optimization platform, NetsPresso. This intuitive interface simplifies the AI optimization process with advanced layer-wise analysis and automated quantization.
How to Run Coding Assistants for Free on RTX AI PCs and Workstations
This blog post was originally published at NVIDIA’s website. It is reprinted here with the permission of NVIDIA. AI-powered copilots deliver real-time assistance for projects from academic projects to production code — and are optimized for RTX AI PCs. Coding assistants or copilots — AI-powered assistants that can suggest, explain and debug code — are
Microchip Technology Demonstration of Real-time Object and Facial Recognition with Edge AI Platforms
Swapna Guramani, Applications Engineer for Microchip Technology, demonstrates the company’s latest edge AI and vision technologies and products at the 2025 Embedded Vision Summit. Specifically, Guramani demonstrates her company’s latest AI/ML capabilities in action: real-time object recognition using the SAMA7G54 32-bit MPU running Edge Impulse’s FOMO model, and facial recognition powered by TensorFlow Lite’s Mobile
