Intra prediction with 3-tap filters for lossless and lossy video coding

---

Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Elektrik ve Elektronik Mühendisliği Bölümü, Türkiye

Tezin Onay Tarihi: 2016

Öğrenci: SAEED RANJBAR ALVAR

Danışman: FATİH KAMIŞLI

Özet:

Video coders are primarily designed for lossy compression. The basic steps in modern lossy video compression are block-based spatial or temporal prediction, transformation of the prediction error block, quantization of the transform coefficients and entropy coding of the quantized coefficients together with other side information. In some cases, this lossy coding architecture may not be efficient for compression. For example, when lossless video compression is desirable, the transform and quantization steps are skipped. Or in lossy compression of synthetic video content (such as animations), the transform may be skipped for some of the blocks and the prediction error is quantized and entropy coded in those blocks. In these cases, the block-based spatial prediction (called intra prediction) cannot sufficiently decorrelate the pixels by itself and large prediction errors become more frequent. For the cases where the transform is skipped, the block-based prediction can be replaced with a more accurate pixel-by-pixel prediction since the original/reconstructed neighboring pixels inside the block will be readily available due to the lack of transform. This thesis explores pixel-by-pixel prediction methods based on 3-tap filtering which use three neighboring pixels for prediction according to a two-dimensional correlation model. Two of the proposed methods are designed for lossless intra coding, one with offline determined prediction weights and the other with online determined adaptive weights. The third proposed method uses the 3-tap filtering method for the transform skipped blocks in lossy intra coding. The proposed methods are implemented within the HEVC reference software and the experimental results indicate that the pixel-by-pixel spatial prediction method based on 3-tap filtering can improve the compression efficiency for both lossless and lossy coding.