Visidon Presents Enhanced Object Detection at Embedded World 2024

Visidon has been invited to present on empowering AI to enhance object detection in extreme low-light environments at this year’s Embedded World Conference in Nuremberg, Germany, taking place from April 9th to 11th.

The presentation by Mr. Sami Pietilä, Head of Embedded Video at Visidon, will be held on April 9th at 11 AM. Additionally, we will be showcasing a real-time live demo of low-light denoising and detection on the H15 camera device at our partner Hailo’s stand in Hall 1, stand 126.

More about the topic:

Empowering AI & Embedded Vision: Enhancing Object Detection in Extreme Low-Light Environments

This presentation delves into the challenges and solutions for detecting visual objects in extreme low-light conditions, specifically in environments with less than 1 Lux of light.

In low-light conditions, noise can make it difficult for computer vision algorithms to accurately detect and recognize objects. This is because the noise can introduce false positives (detecting objects that aren’t there) or false negatives (missing objects that are present). In this presentation, we discuss the latest advancements in Embedded AI, such as convolutional neural networks for noise reduction. We also address the concern of motion blur and subsequently delve into the compromise between reducing exposure times and elevating noise levels.

In our presentation, we discuss the delicate balance between visual quality and object recognition accuracy: when it comes to object detection and recognition, it may not always align with the requirements of visual quality enhancement. Another crucial aspect addressed is the utilization of 4K@30fps video enhancement on the edge, enabling real-time, high-resolution video processing enhancing object detection accuracy.

By presenting real-world application samples with Edge AI processor, we demonstrate how CNN-based low-light enhancement and noise reduction impact the performance of cameras in extreme low-light conditions, where even the premium cameras´ detectors might struggle.

Visiting Embedded World:

We invite you to visit Embedded World Conference! Use the e-code “ew24web” for free exhibition access. Separate tickets must be booked for the conferences –  register here!

Contact Mr. Pietilä to schedule a meeting at the event: Sami Pietilä, [email protected]

Here you’ll find a wealth of practical technical insights and expert advice to help you bring AI and visual intelligence into your products without flying blind.

Contact

Address

Berkeley Design Technology, Inc.
PO Box #4446
Walnut Creek, CA 94596

Phone
Phone: +1 (925) 954-1411
Scroll to Top