Intra-frame
In video compression, an intra-frame or key frame ( intra-coded frame , I-frame or key frame ) is an individual image ( frame ) that can be decoded independently of other individual images. In the inter-frame encoding multiple frames are a group of pictures (English Group of Pictures, GOP summarized). In addition to intra-frames, there are inter-frames that relate to neighboring frames. There are two types of inter-frames, P- and B-pictures . Compared to intra-frames , inter-frames save storage space , since referencing to image areas of other frames means that no direct image data, but only reference information and motion vectors are saved.
Often at least a minimum frequency of intra-frames is enforced in the coded video in order to keep the amount of random access to images to be decoded (for example also entry into a stream ) within a safe framework. With MPEG-1 and -2 there is typically one intra-picture every half second, with MPEG-4 often one every ten seconds. From H.264 , a group of pictures can contain several I-pictures.
Mainly for use with video cameras there are video codecs that are limited to the use of intra-images:
They can be coded with little computational effort, their image quality is independent of image movement with random access, but they require a comparatively large amount of storage space, since no consideration is given to similarities to previous or subsequent frames.
The use of neural networks for intra-frame prediction is a research area.
Individual evidence
- ↑ Chair for Multimedia Communication and Signal Processing: Image and Video Coding. Retrieved February 28, 2020 .