Software for Embedded Vision

Efficient LLaMA-3.2-Vision by Trimming Cross-attended Visual Features
This blog post was originally published at Nota AI’s website. It is reprinted here with the permission of Nota AI. Our method, Trimmed-Llama, reduces the key-value cache (KV cache) and latency of cross-attention-based Large Vision Language Models (LVLMs) without sacrificing performance. We identify sparsity in LVLM cross-attention maps, showing a consistent layer-wise pattern where most

D3 Embedded to Showcase Robotic Perception and Generative AI Solutions at the Embedded Vision Summit
D3 Embedded to demonstrate real-time solutions integrating cameras and 3D sensors, robust connectivity, embedded processing, and generative AI at the Embedded Vision Summit. Rochester, NY – May 15, 2025 – D3 Embedded announced today it will exhibit at the 2025 Embedded Vision Summit, the premier event for practical, deployable computer vision and AI, for product

Deploying an Efficient Vision-Language Model on Mobile Devices
This blog post was originally published at Nota AI’s website. It is reprinted here with the permission of Nota AI. Recent large language models (LLMs) have demonstrated unprecedented performance in a variety of natural language processing (NLP) tasks. Thanks to their versatile language processing capabilities, it has become possible to develop various NLP applications that

Qualcomm AI Inference Suite: Getting Started is Easy
This blog post was originally published at Qualcomm’s website. It is reprinted here with the permission of Qualcomm. Once you have a key, it is simply a matter of choosing how to connect to the inference endpoint. If you are most comfortable with Python, an SDK is provided along with documentation so that you can

LM Studio Accelerates LLM Performance With NVIDIA GeForce RTX GPUs and CUDA 12.8
This blog post was originally published at NVIDIA’s website. It is reprinted here with the permission of NVIDIA. Latest release of the desktop application brings enhanced dev tools and model controls, as well as better performance for RTX GPUs. As AI use cases continue to expand — from document summarization to custom software agents —

AI Agents, Explained: Use Cases, Potential and Limitations
This blog post was originally published at Tryolabs’ website. It is reprinted here with the permission of Tryolabs. AI agents have taken center stage in tech conversations over the past year. Bold claims swirl about how they’ll reinvent workflows, slash costs, and even replace human teams. But with so much hype in the air, it’s

Advancing Generative AI at the Edge During CES 2025
This blog post was originally published at Ambarella’s website. It is reprinted here with the permission of Ambarella. For this year’s CES, our theme was Your GenAI Edge—highlighting how Ambarella’s AI SoCs continue to redefine what’s possible with generative AI at the edge. Building on last year’s edge GenAI demos, we debuted a new 25-stream,

Optimizing Transformer-based Diffusion Models for Video Generation with NVIDIA TensorRT
This article was originally published at NVIDIA’s website. It is reprinted here with the permission of NVIDIA. State-of-the-art image diffusion models take tens of seconds to process a single image. This makes video diffusion even more challenging, requiring significant computational resources and high costs. By leveraging the latest FP8 quantization features on NVIDIA Hopper GPUs

Imaging Trends in 2025: Shaping the Future of Security and Surveillance
This blog post was originally published at Visidon’s website. It is reprinted here with the permission of Visidon. A couple of weeks ago, we had the opportunity to exhibit at ISC West 2025 in Las Vegas—the flagship global event for security and surveillance professionals. As always, this show served as a crystal ball for the

Enable Pose Detection on Snapdragon X Elite: Step-by-step Tutorial
This article was originally published at Qualcomm’s website. It is reprinted here with the permission of Qualcomm. I know why you’re here; you’ve decided to buy your first device with Snapdragon X Elite processor, awesome choice! You now ventured over to Qualcomm AI Hub, grabbed a model and excitedly watched as it downloaded. “Hmmm okay…

R²D²: Adapting Dexterous Robots with NVIDIA Research Workflows and Models
This blog post was originally published at NVIDIA’s website. It is reprinted here with the permission of NVIDIA. Robotic arms are used today for assembly, packaging, inspection, and many more applications. However, they are still preprogrammed to perform specific and often repetitive tasks. To meet the increasing need for adaptability in most environments, perceptive arms

Efficient Optical Character Recognition (OCR)
This blog post was originally published at Basler’s website. It is reprinted here with the permission of Basler. Discover our easy and efficient way of implementing an OCR solution, with chip and IC trays as the particular use case example, where challenging characters are deciphered correctly every time. The solution overcomes low contrast and variable

Automotive OEMs Integrating AI into In-cabin Sensing
Integration of AI into In-Cabin Sensing – Automotive OEMs Making Full Use of Software to Enable Advanced Features As regulatory frameworks such as the EU’s Advanced Driver Distraction Warning (ADDW) near enforcement, automotive OEMs are increasingly integrating in-cabin sensing technologies with software-defined vehicle architectures. This shift, evident at CES 2025, marks a pivotal moment in

ELD: Introducing a New Open-source Embedded Linker Tool for Embedded Systems
This blog post was originally published at Qualcomm’s website. It is reprinted here with the permission of Qualcomm. At Qualcomm Technologies, Inc., embedded linkers play a crucial role in our software stack. While many linkers work well on traditional platforms, they often fall short when it comes to embedded systems. Embedded projects have unique requirements,

North America and Europe to Account for 17 Million Video Telematics Systems in Use by 2029
For more information, visit https://www.berginsight.com/the-video-telematics-market. The integration of cameras to enable various video-based solutions in commercial vehicle environments is one of the most apparent trends in the fleet telematics sector today. Berg Insight’s definition of video telematics includes a broad range of camera-based solutions deployed in commercial vehicle fleets either as standalone applications or as

Using AI to Better Understand the Ocean
This blog post was originally published at NVIDIA’s website. It is reprinted here with the permission of NVIDIA. Humans know more about deep space than we know about Earth’s deepest oceans. But scientists have plans to change that—with the help of AI. “We have better maps of Mars than we do of our own exclusive

SWIR Vision Systems in Agricultural Production
This blog post was originally published at Basler’s website. It is reprinted here with the permission of Basler. Improved produce inspection through short-wave infrared light Ensuring the quality of fruits and vegetables such as apples or potatoes is crucial to meet market standards and consumer expectations. Traditional inspection methods are often based only on visual

Qualcomm Dragonwing Intelligent Video Suite Modernizes Video Management with Generative AI at Its Core
This blog post was originally published at Qualcomm’s website. It is reprinted here with the permission of Qualcomm. Video cameras generate a lot of data. Companies that use a video management system (VMS) are left wanting to get more value out of all the video data they generate, enabling them to take the actions that