LipSync: Audio-Video Synchronization Detection – Intel Chip Chat Network Insights – Episode 241

November 26th, 2019 |
Image for FaceBook

 
Share this post:
Facebook | Twitter | Google+ | LinkedIn | Pinterest | Reddit | Email
 
This post can be linked to directly with the following short URL:


 
The audio player code can be copied in different sizes:
144p, 240p, 360p, 480p, 540p, Other


 
The audio player code can be used without the image as follows:


 
This audio file can be linked to by copying the following URL:


 
Right/Ctrl-click to download the audio file.
 
Subscribe:
Connected Social Media - iTunes | Spotify | Google | Stitcher | TuneIn | Twitter | RSS Feed | Email
Intel - iTunes | Spotify | RSS Feed | Email
Intel Chip Chat - iTunes | Spotify | RSS Feed | Email
Intel Chip Chat: Network Insights - iTunes | Spotify | RSS Feed | Email
 

MulticoreWare’s LipSync technology uses Deep Neural Networks to “watch” and “listen” to videos and determine if content is correctly synchronized. Trained classifiers are used to find and match human faces to human speech, allowing LipSync to detect errors that are not found by current automated video quality control systems.

MulticoreWare was founded by a team of researchers that wanted a better way to program for heterogeneous architectures. With the advent of GPGPUs and the increasing prevalence of multi-core, multi-architecture platforms, our clients were struggling with the difficulties of using these platforms efficiently. Expertise covers HPC and Cloud Computing, GPGPUs, DSPs, FPGAs, and mobile and embedded platforms.

Vinod Kannan is Director of Engineering – Machine Learning Solutions for MulticoreWare, and joins Intel Chip Chat Network Insights in this archive of a live cast interview from the Visual Cloud conference and IBC Show 2019, in Amsterdam, Netherlands.

In this interview, Vinod describes MulticoreWare’s involvement in the upcoming Intel Select Solution for Media Analytics. Hear about the solution components, Intel Xeon Scalable processors, also the Intel Movidius Myriad X vision processing unit (VPU) to provide real-time video and Deep Learning processing designed for the network edge through the Intel Visual Cloud Accelerator Card—Analytics (Intel VCAC-A). All combined, this provides an amazing opportunity for underlying hardware, the OpenVINO toolkit and the LipSync solution to deliver a compelling solution to the marketplace.

For more information visit:
multicorewareinc.com

The Intel Select Solutions for Media Analytics give OEMs, ISVs, and CommSPs the means to accelerate delivery of products and services based on powerful insights into the contents of video streams. Pre-validated stacks of hardware and software provide a head start for solution providers, enabling them to forgo complex platform questions so they can focus on value-added IP. The future of media analytics has arrived.

Learn more:
builders.intel.com/intelselectsolutions/intel-select-solutions-for-media-analytics
IBC and Intel collaborated on the ‘Visual Cloud Conference’ that brought together thought leaders across the industry to drive dialogue and advancements around the future of streaming media and other visual cloud services.

Explore Visual Cloud demos and presentations from the event here:
intel.com/visualcloud

Learn more about Intel’s work in network transformation, 5G and edge computing at:
networkbuilders.intel.com

Tags: , , , , , , , , , , , , , , , , , , , , , ,
 
Posted in: Artificial Intelligence, Audio Podcast, Cloud Computing, Intel, Intel Chip Chat, Intel Chip Chat: Network Insights
 


 


Learn more with Intel IT’s expert on IT Cloud Computing, Das Kamhout on:
Intel.com, LinkedIn, Google+, or follow him on Twitter.