Development of a solution to asses the quality and to optimize AI-based video codecs - QC-232Desired discipline(s): Engineering - computer / electrical, Engineering, Computer science, Mathematical Sciences, Mathematics
Company: Avid Technology, Inc.
Project Length: 4 to 6 months
Preferred start date: 01/01/2020
Language requirement: EnglishEnglish with some French proficiencyFrench with some English proficiency
Location(s): Montreal, QC, Canada
No. of positions: 2
About the company:
Thirty years ago, Avid broke new ground by reimagining content creation. Our revolutionary nonlinear editor was the first to digitize video content. It redefined the media industry and is still the gold standard.
Today, Avid is reshaping the entire media value chain with powerful technology for creating, managing, storing, distributing and monetizing film, television, and music. Our tools and platforms empower more than a million users and thousands of media enterprises to tell powerful stories and build better businesses.
With enhanced collaboration, advanced automation, end-to-end integration, and workflow orchestration, Avid today helps aspiring artists, creative professionals, production teams and media enterprises to thrive in the digital era.
Please describe the project.:
The current project aims to assess and optimize AI-based video compression algorithms.
- Assess the quality of AI-based video codecs from an industry perspective
- Determine if the codec can run on off-the-shelf hardware
- Evaluation of optimization techniques for AI-based codecs (i.e. vectorization, SIMD, GPU, etc.)
The skills for this project include:
- Image quality evaluation (i.e. calculating PSNR, SSIM, etc.)
- Benchmarking the performance of CPU and GPU utilization
- Machine learning, with a focus on image processing
- Knowledge of Python, C/C++
- Development and testing under the following environments: Windows, Mac OS, Linux
- Usage of existing libraries: TensorFlow, Keras, OpenCV, Numpy, Intel MKL