Efficient Extraction of motion vectors from h264/h265 streams



A video encoder is a system that is able to transform a raw video (a sequence of uncompressed frames) to a more transferrable and storable format. Conversely a video decoder is a system that transforms that streams back to images. The pair (encoder, decoder) is usually called a video codec.

In order to achieve an high level of compression without loosing to much quality, some of the most used video encoders (e.g. AVC and HEVC) decompose each frame in tiles and if most of these will stay unchanged only give  a motion estimation of the changing ones.

These motion vectors may be very usefull for video analysis, so we need an efficient way to extract them trying to exploit hardware optimisation.

Planned Activities

The first part of the thesis will be about research state of the art codec implementation (for codecs that use motion vectors).

After chosing one of those codecs the students should write and efficient library that given a videostream will extract motion vectors on a chosen computer architecture.

Addfor will provide a widespread choice of hardware solutions.

Duration of this Project: 5-6 months.

Who we’re looking for

Students that are about to get their master degree in: Computer Science, Computer Engineering.

Skills: Low level programming,  C/C++, Cuda programming, strong knowledge of computer architectures.

Check these links before moving on

Contact Us