Lists

Tuesday, January 16, 2024

OPPO, AlpsenTek and Qualcomm collaboration on "Hybrid Vision Sensing"

OPPO, AlpsenTek and Qualcomm Boost AI Motion, Image Quality For Mobile Applications

Jan.11,2024,Las Vegas,USA—OPPO, AlpsenTek and Qualcomm Technologies, Inc. have teamed up with the goal of enhancing innovative Hybrid Vision Sensing (HVS) technology, to better extract valuable motion and image data to enhance picture quality for mobile phone applications.

OPPO and AlpsenTek will collaborate to pioneer the use of Hybrid Vision Sensing technologies, developing a data processing chain to collect relevant camera information to help enhance picture quality and allow for deblurring, augmented resolution, and slow-motion reconstruction, as well as other features required for machine sensing. This will be accomplished by leveraging Snapdragon® Mobile Platforms from Qualcomm Technologies.

“The HVS solution, with the support of hardware and algorithms, significantly enhances the capacities of smartphone cameras”, said Judd Heape VP, Product Management at Qualcomm Technologies, Inc. “We are pleased to contribute to the optimization of this new technology on our Snapdragon platforms – which will help consumers to get the best performance from their smartphone cameras, and capture what’s most precious to them.”

The Image product director, Mr. Xuan ZHANG, from OPPO commented: “Over the years, we have conducted extensive research in new sensor technologies, with a particular focus on HVS (Hybrid Vision System) technology. We have engaged in substantial collaborative developments with AlpsenTek and Qualcomm, involving numerous iterations in both chip design and algorithms. Our confidence in the potential of this technology has driven us to invest time and effort into refining it collaboratively, with the ultimate goal of pushing it towards the application on OPPO’s HyperTone Camera System.” 

Motion information is crucial in photography and machine vision. Traditional image sensors collapse motion information within a period (i.e. the exposure) into a single image. This leads to motion blurs and loss of valuable motion data essential for image/video processing and machine vision algorithms.

Effectively obtaining high-fidelity motion information with a vision sensor is a top demand across various fields today. Current solutions based on conventional image sensors often rely on increasing the frame rate, which is expensive and impractical for many applications. High frame rates lead to a significant amount of data (much of it redundant) and short shutter durations, causing high system resource usage, low efficiency, and poor adaptation to lighting conditions for high-frame-rate cameras.

Event-based Vision Sensing (EVS) is an imaging technology that continuously records change/motion information through its shutter-free mechanism. It provides motion information with high time resolution and lower cost machine vision. With an in-pixel processing chain featuring logarithm amplification, EVS achieves a balance between high frame rate, high dynamic range, and low data redundancy for recording motion information.

However, EVS sensors often lack critical static pictorial information that is needed for many machine vision applications. It typically works alongside a separate traditional image sensor (RGB) to compensate for this drawback, introducing challenges in cost, system complexity, and image registration between the two types of images (EVS and RGB), offsetting many of EVS's advantages.
AlpsenTek's Hybrid Vision Sensing (HVS) technology, introduced in 2019, combines EVS and conventional imaging technology into a single sensor. The ALPIX® sensor from AlpsenTek simultaneously outputs high-quality RGB images and EVS data stream, providing a cost-effective and algorithm-friendly solution for capturing images with embedded motion information.

Jian Deng, Founder and CEO of AlpsenTek, stated, "In the current landscape of vision sensors, there is a growing expectation for more than just 2D RGB information; sensors are now anticipated to provide additional data, such as distance, spectrum, and motion. Collaborating with OPPO and Qualcomm, we collectively designed the ALPIX-Eiger® to seamlessly integrate into mobile phone applications. Considered an enhanced RGB image sensor, it boasts image quality comparable to leading mobile sensors on the market, while introducing the added functionality of EVS. Witnessing the process of bringing our technology from conception to product brings us immense excitement."

Deng further emphasized, "It's important to recognize that what truly changes the world is not the technology itself but the products that it enables. Our passion lies in bringing Hybrid Vision Sensing (HVS) into the hands of everyone. This commitment has been our driving force from the very beginning. We look forward to fruitful outcomes from this collaboration”.

Jian Deng, Founder and CEO of AlpsenTek, stated, "In the current landscape of vision sensors, there is a growing expectation for more than just 2D RGB information; sensors are now anticipated to provide additional data, such as distance, spectrum, and motion. Collaborating with OPPO and Qualcomm, we collectively designed the ALPIX-Eiger® to seamlessly integrate into mobile phone applications. Considered an enhanced RGB image sensor, it boasts image quality comparable to leading mobile sensors on the market, while introducing the added functionality of EVS. Witnessing the process of bringing our technology from conception to product brings us immense excitement."

Deng further emphasized, "It's important to recognize that what truly changes the world is not the technology itself but the products that it enables. Our passion lies in bringing Hybrid Vision Sensing (HVS) into the hands of everyone. This commitment has been our driving force from the very beginning. We look forward to fruitful outcomes from this collaboration”.

This news was also featured on EETimes: https://www.eetimes.com/oppo-alpsentek-and-qualcomm-boost-ai-motion-image-quality-for-mobile-applications/

6 comments:

  1. a lot of money to wast.

    ReplyDelete
  2. is there a high volume application for event image sensors like smartphones yet? have not heard about a smartphone usecase before

    ReplyDelete
  3. This looks like simulated data.

    ReplyDelete
  4. It looks similar to OVT's ISSCC 2023 "15MPixel CIS + 1MPixel EVS" idea

    ReplyDelete
  5. It is also interesting from a automotive point of view for ADAS and autonomous driving to get info on motion and/or reduce motion blur. Especially side cameras are really blurry.
    However, there are rolling shutter sensors used.

    ReplyDelete

All comments are moderated to avoid spam and personal attacks.