Fast H.264/AVC to HEVC transcoding based on machine learning

Peixoto, Eduardo; Macchiavello, Bruno; Queiroz, Ricardo L. de; Hung, Edson M.

doi:10.1109/its.2014.6947999

Cited by 17 publications

(6 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In literature survey, this research work have spread light on various authors [1][2][3][4] published various results of different transcoding techniques for HEVC. Transcoding, which is done on the basis of mode mapping [1] is implemented to get more stable results over wide range of sequences.…”

Section: Literature Surveymentioning

confidence: 99%

See 1 more Smart Citation

Survey of Various Advanced Video Compression and Encryption Techniques

Dhanalakshmi¹

2019

IJRASET

View full text Add to dashboard Cite

In recent decades, video compression has become the primary focus of intensive studies. Besides there is a great development in video compression technique in these days; and this technique is used in large number of applications. Video compression techniques is used to reduce and remove the less important video data and so the digital video data will be sent effectively and received over a network and stored on computer disks. In this paper, the research is set out with an idea about various techniques available for video compression. Nowadays, numerous video compression techniques were introduced. Those are with significant compression techniques, a vita land the reduced file will be gained without any loss in effect in the visual quality. H.265/HEVC exhibits higher improvement in coding performance over its predecessors. These are both VCEG and MPEG articulating the next generation norms. The overview of the advancement of these next-generation video coding initiatives and current upcoming video coding techniques was provided in this research work. Several papers and research materials were published related to these subject and some of these are taken for this research study. This survey is conducted to define the different fundamental methods available for video compression; video compression techniques application strategies ; different generations; the upcoming video compression techniques and trends; various issues parallel with the emerging technologies.

show abstract

Section: Literature Surveymentioning

confidence: 99%

“…As there are very large numbers of modes in HEVC, mode mapping is a significant technique of transcoding. Higher speed for transcoding will be achieved through machine learning technique [2]. For complexity reduction, the new codec based on complexity scalable technique is introduced [3].…”

Section: Literature Surveymentioning

confidence: 99%

Survey of Various Advanced Video Compression and Encryption Techniques

Dhanalakshmi¹

2019

IJRASET

View full text Add to dashboard Cite

show abstract

“…Thus, the overall selected percentage of modes by the proposed method at 32×32, 16×16 and 8×8 coding depth level for each QPs (from 40 to 20) seem to be more consistent. This strategy influences the concept of modes transcending [29]. Furthermore, the selection of coding depth level 2 or higher level modes (as stated earlier) is assumed to be an indication of high motion dominating region.…”

Section: Mode Analysis Of Coding Depth Levelsmentioning

confidence: 99%

A novel motion classification based intermode selection strategy for HEVC performance improvement

2016

View full text Add to dashboard Cite

“…A fast MB mode decision scheme based on support vector machines is proposed for video transcoder in [10]. In 2014, Peixoto proposed an H.264/AVC to HEVC transcoding algorithm based on ML [11]. In [12], Peixoto proposed a transcoding algorithm based on a dynamic threshold and content modelling; in this algorithm, the first k frames of a sequence are used for training linear discriminant functions, so that the transcoder can learn the classifier for subsequent frames.…”

Section: Introductionmentioning

confidence: 99%

“…Thus, it is difficult to attain a satisfactory tradeoff between complexity and RD performance for different sequences. Additionally, the ML tools in [11, 13, 14] are all based on decision trees; the classifier accuracy of decision trees is below that of the tree‐augmented Naive Bayesian (NB) classifier (TAN) [15], which is used to train the classifier in this paper.…”

Section: Introductionmentioning

confidence: 99%