基于AVS2的色度擴(kuò)展視頻編碼的設(shè)計(jì)和實(shí)現(xiàn)

王淑慧

doi:10.11999/JEIT180154

基于AVS2的色度擴(kuò)展視頻編碼的設(shè)計(jì)和實(shí)現(xiàn)

doi: 10.11999/JEIT180154 cstr: 32379.14.JEIT180154

王淑慧^,

同濟(jì)大學(xué)電子與信息工程學(xué)院 ??上海 ??201804

基金項(xiàng)目: 國(guó)家自然科學(xué)基金(61601200, 61871289)，上海市自然科學(xué)基金(18ZR1440600)

詳細(xì)信息

作者簡(jiǎn)介:
王淑慧：女，1973年生，副研究員，研究方向?yàn)橐曨l編碼算法及SOC實(shí)現(xiàn)

通訊作者:
王淑慧　 shw@#edu.cn

中圖分類號(hào): TN919.8
計(jì)量
- 文章訪問數(shù): 2013
- HTML全文瀏覽量: 796
- PDF下載量: 32
- 被引次數(shù): 0
出版歷程
- 收稿日期: 2018-02-06
- 修回日期: 2018-09-14
- 網(wǎng)絡(luò)出版日期: 2018-09-18
- 刊出日期: 2018-12-01

Design and Implementation for Chroma Extensions Video Coding Based on AVS2 Platform

Shuhui WANG^,

College of Electronics and Information Engineering, Tongji University, Shanghai 201804, China

Funds: The National Natural Science Foundation of China (61601200, 61871289), Shanghai Natural Science Foundation (18ZR1440600)

摘要

摘要: 色度擴(kuò)展視頻編碼是當(dāng)前視頻編碼領(lǐng)域的一個(gè)熱點(diǎn)研究課題。該文提出基于AVS2平臺(tái)的色度擴(kuò)展視頻幀內(nèi)編碼的實(shí)現(xiàn)方案。仿444/422編碼方案通過將輸入圖像中的色度分量下采樣后，使用原有的420方式進(jìn)行編碼，以實(shí)現(xiàn)444/422編碼。進(jìn)一步，該文將幀內(nèi)預(yù)測(cè)及環(huán)路濾波等編碼模塊無縫擴(kuò)展到相應(yīng)的444/422格式，實(shí)現(xiàn)444/422幀內(nèi)預(yù)測(cè)編碼。實(shí)驗(yàn)結(jié)果表明，對(duì)444格式和422格式序列，在高碼率的情況下444/422幀內(nèi)預(yù)測(cè)編碼與仿444/422編碼相比，U/V平均BD-rate的減少分別為31.44%/31.72%和18.85%/19.3%，而Y分量平均BD-rate的增加僅為0.5%。其中422色度幀內(nèi)預(yù)測(cè)過程的算法優(yōu)化減少Y/U/V BD-rate最高可達(dá)5.66%。與HEVC RExt編碼相比，在低碼率時(shí)，444/422幀內(nèi)預(yù)測(cè)編碼取得了更好或相近的編碼性能。
- 色度擴(kuò)展視頻編碼 /
- 色度子采樣 /
- AVS2 /
- 幀內(nèi)預(yù)測(cè)
Abstract: Chroma extensions video coding is a hot topic in the field of video coding. Chroma extensions video coding scheme based on AVS2 platform is proposed. The most direct solution is pseudo444/422 coding. In this method, chroma component in the input image is down sampled by averaging adjacent samples. The core coding modules are still 420 coding. Further, this paper seamlessly extends intra prediction and loop filter to the 444/422 chroma format to implement 444/422 intra prediction coding. The experimental results show that compared with pseudo444/422 coding, in the case of high bit rate, the average U/V BD-rate saving is 31.44%/31.72% and 18.85%/19.30% for 444 and 422 test sequences respectively, with negligible increase of Y BD-rate (0.5% on average). The modification of the 422 chroma intra prediction algorithm achieves up to 5.66% Y/U/V BD-rate reduction. 444/422 intra prediction coding provides similar or better coding performance than HEVC RExt coding at low bitrates.
- Chroma extensions video coding /
- Chroma subsampling /
- AVS2 /
- Intra prediction

HTML全文

圖 1 HEVC RExt中的全444/422編碼架構(gòu)

下載: 全尺寸圖片幻燈片

圖 2 P444/422編碼架構(gòu)

下載: 全尺寸圖片幻燈片

圖 3 IP444/422編碼架構(gòu)

下載: 全尺寸圖片幻燈片

圖 4 不同色度格式對(duì)應(yīng)的幀內(nèi)預(yù)測(cè)塊劃分方式

下載: 全尺寸圖片幻燈片

圖 5 Bilinear模式預(yù)測(cè)

下載: 全尺寸圖片幻燈片

圖 6 角度模式示意圖

下載: 全尺寸圖片幻燈片

表 1 8個(gè)測(cè)試序列

序號(hào)	序列名	短名	分辨率	比特?cái)?shù)/樣點(diǎn)	幀率
1	iceAerial_3840×2160_444(422)_10 bit	ARL	3840×2160	10	30
2	iceRiver_3840×2160_444(422)_10 bit	RVR	3840×2160	10	30
3	iceRock_3840×2160_444(422)_10 bit	RCK	3840×2160	10	30
4	iceRock2_3840×2160_444(422)_10 bit	RK2	3840×2160	10	30
5	Bubbles_4096×2160_24_10bit_444(422)	BBE10b	4096×2160	10	24
6	Traffic_2560×1600_30_10 bit_444(422)_crop	TFC10b	2560×1600	10	30
7	Bubbles_4096×2160_24_12 bit_444(422)	BBE12b	4096×2160	12	24
8	Traffic_2560×1600_30_12 bit_444(422)_crop	TFC12b	2560×1600	12	30

下載: 導(dǎo)出CSV

表 2 兩組測(cè)試QP

	傳統(tǒng)QP組	超高質(zhì)量QP組
P444/422及IP444/422編碼器, 10-bit序列	41, 48, 54, 61	14, 21, 28, 34
P444/422及IP444/422編碼器, 12-bit序列	57, 64, 70, 77	30, 37, 44, 50
HM16.6編碼器	22, 27, 32, 37	2, 7, 12, 17

下載: 導(dǎo)出CSV

表 3 IP444/422與P444/422實(shí)驗(yàn)結(jié)果比較(%)

	序列名	QP=14, 21, 28, 34 or 30, 37, 44, 50			QP=41, 48, 54, 61 or 57, 64, 70, 77
	序列名	Y	U	V	Y	U	V
YUV444	ARL	–0.59	–71.07	–47.01	0.08	–4.42	–2.31
	RVR	–0.21	–25.13	–39.49	0.43	4.49	1.36
	RCK	0.18	–16.67	–14.77	0.43	7.47	7.88
	RK2	0.21	–27.59	–18.88	0.43	2.29	0.51
	BBE10b	0.17	–48.05	–38.37	0.13	–22.79	–15.33
	TFC10b	1.85	–27.33	–55.52	0.53	–6.05	–9.89
	BBE12b	0.24	–28.43	–18.69	–0.14	–15.12	–8.67
	TFC12b	2.12	–7.27	–21.05	0.12	–3.76	–5.52
444平均		0.50	–31.44	–31.72	0.25	–4.74	–4.00
YUV422	ARL	0.08	–32.45	–21.12	5.45	–3.48	–1.82
	RVR	0.58	–18.22	–29.51	3.58	4.79	1.99
	RCK	0.65	–11.52	–9.44	3.94	8.96	8.74
	RK2	0.71	–21.95	–13.92	3.89	1.92	1.75
	BBE10b	0.06	–19.36	–18.34	1.27	–2.63	–0.68
	TFC10b	0.81	–21.35	–38.49	3.46	–2.43	–5.66
	BBE12b	0.43	–8.77	–6.16	0.18	–2.98	–1.37
	TFC12b	0.90	–17.14	–17.42	0.30	–1.69	–2.81
422平均	平均	0.53	–18.85	–19.30	2.76	0.31	0.02

下載: 導(dǎo)出CSV

表 4 IP444/422編碼與HM16.6實(shí)驗(yàn)結(jié)果比較(%)

	序列名	QP=14, 21, 28, 34 or 30, 37, 44, 50				QP=41, 48, 54, 61 or 57, 64, 70, 77
	序列名	Y	U	V	Cpsnr	Y	U	V	Cpsnr
YUV444	ARL	–8.91	71.73	65.46	–2.32	–13.07	64.85	58.25	–5.98
	RVR	–8.66	87.33	117.26	–1.61	–6.25	100.29	97.44	–0.39
	RCK	–6.27	65.96	68.14	–3.26	–9.45	41.87	40.87	–6.92
	RK2	–6.60	78.85	82.33	–3.04	–8.97	41.30	36.70	–6.84
	BBE10b	–21.38	144.38	100.00	1.08	–34.36	99.20	101.21	–14.43
	TFC10b	–33.59	100.00	100.00	71.28	–20.12	83.45	68.89	–8.13
	BBE12b	–20.58	100.00	100.00	12.25	–25.99	138.27	125.42	–2.67
	TFC12b	–32.97	100.00	100.00	66.38	–12.85	84.05	81.61	–1.69
444平均		–17.37	93.53	91.65	17.60	–16.38	81.66	76.30	–5.88
YUV422	ARL	–5.32	11.11	12.37	–3.01	–6.08	7.18	5.96	–4.22
	RVR	–5.67	26.32	37.23	–1.82	–2.27	13.82	17.62	–0.31
	RCK	–3.18	25.05	24.41	–1.45	–4.34	3.84	2.76	–3.69
	RK2	–3.10	29.96	29.46	–1.01	–3.66	1.06	–4.21	–3.51
	BBE10b	–11.98	23.52	23.79	–3.75	–16.58	18.07	25.56	–9.42
	TFC10b	–16.17	100.00	77.21	23.00	–6.84	18.05	14.23	–2.53
	BBE12b	–10.98	49.59	54.40	2.84	–7.30	31.64	36.23	0.66
	TFC12b	–15.42	100.00	126.25	22.01	–2.16	20.82	21.55	1.66
422平均	平均	–8.98	45.69	48.14	4.60	–6.15	14.31	14.96	–2.67

下載: 導(dǎo)出CSV

表 5 422格式色度幀內(nèi)預(yù)測(cè)修改的實(shí)驗(yàn)結(jié)果(%)

序列名	QP=14, 21, 28, 34 or 30, 37, 44, 50			QP=41, 48, 54, 61 or 57, 64, 70, 77
序列名	Y	U	V	Y	U	V
ARL	–1.68	–1.61	–1.73%	–3.24	–3.02	–2.73
RVR	–1.96	–1.01	–1.91	–2.91	–2.41	–2.48
RCK	–1.57	–0.17	0.08	–2.82	–5.30	–5.66
RK2	–1.43	–1.22	0.09	–2.40	–2.50	–2.74
BBE10b	–0.86	–0.82	–0.70	–2.48	–3.53	–3.37
TFC10b	–1.05	–2.60	–1.25	–1.71	–2.60	–1.22
BBE12b	–0.80	–2.54	–2.19	–5.44	–2.13	–0.77
TFC12b	–1.14	–3.70	–4.47	–2.90	–2.20	–1.58
平均	–1.31	–1.71	–1.51	–2.99	–2.96	–2.57

下載: 導(dǎo)出CSV

參考文獻(xiàn)(18)

FRAN?OIS E, RUSANOVSKYY D, YIN P, et al. Suggested new draft text of signalling, backward compatibility and display adaptation for HDR/WCG video coding[C]. JCTVC-Z0023, Geneva, Switzerland, 2017: 1–31.

GOMMELET D, ROUMY A, GUILLEMOT C, et al. Gradient-based tone mapping for rate-distortion optimized backward-compatible high dynamic range compression[J]. IEEE Transactions on Image Processing, 2017, 26(12): 5936–5949 doi: 10.1109/TIP.2017.2740159

HENRY F, JUNG J, OUACH A, et al. Stereoscopic 360 video compression with the next generation video codec[C]. JVET-G0064, Torino, Italy, 2017: 1–5.

FLYNN D, MARPE D, NACCARI M, et al. Overview of the range extensions for the HEVC standard: Tools, profiles, and performance[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2016, 26(1): 4–19 doi: 10.1109/TCSVT.2015.2478707

PENG W H, WALLS F, COHEN R A, et al. Overview of screen content video coding: Technologies, standards, and beyond[J]. IEEE Journal on Emerging Selected Topics in Circuits and Systems, 2016, 6(4): 393–408 doi: 10.1109/JETCAS.2016.2608971

林濤, 蔡文婷, 陳先義, 等. 一種高性能低復(fù)雜度的基于串匹配的屏幕圖像無損壓縮算法[J]. 電子與信息學(xué)報(bào), 2017, 39(2): 351–359 doi: 10.11999/JEIT160560

LIN Tao, CAI Wenting, CHEN Xianyi, et al. Lossless compression algorithm based on string matching with high performance and low complexity for screen content coding[J]. Journal of Electronics&Information Technology, 2017, 39(2): 351–359 doi: 10.11999/JEIT160560

ZHAO Liping, ZHOU Kailun, GUO Jing, et al. A universal string matching approach to screen content coding[J]. IEEE Transactions on Multimedia, 2018, 20(4): 796–809 doi: 10.1109/TMM.2017.2758519

ZHU Weijia, ZHANG Kai, AN Jicheng, et al. Inter-palette coding in screen content coding[J]. IEEE Transactions on Broadcasting, 2017, 63(4): 673–679 doi: 10.1109/TBC.2017.2711144

FLYNN D, NACCARI M, ROSEWARNE C, et al. High Efficiency Video Coding (HEVC) range extensions text specification: Draft 7[C], JCTVC-Q1005, Valencia, Spain, 2014.

WIEN M, BARONCINI V, BOYCE J, et al. Preliminary joint call for evidence on video compression with capability beyond HEVC[C], JVET-E1002, Geneva, Switzerland, 2017: 1–9.

AVS N2378. AVS第59次會(huì)議視頻組會(huì)議紀(jì)要[C]. 海口, 中國(guó), 2016.

AVS N2378. Meeting summary of video coding subgroup of the 59th AVS meeting[C]. Haikou, China, 2016.

王淑慧, 林濤, 郭靖, 等. AVS2-P2框架下對(duì)YUV444圖像的編碼[C]. AVS M4082, 海口, 中國(guó), 2016: 1–7.

WANG Shuhui, LIN Tao, Guo Jing, et al. YUV444 image coding in AVS2-P2[C]. AVS M4082, Haikou, China, 2016: 1–7.

王淑慧, 林濤, 郭靖, 等. AVS2-P2框架下對(duì)YUV444和YUV422(簡(jiǎn)稱CExt)圖像的編碼[C]. AVS M4154, 北京, 中國(guó), 2017: 1–10.

WANG Shuhui, LIN Tao, Guo Jing, et al. YUV444 and YUV422 image coding in AVS2-P2[C]. AVS M4154, Beijing, China, 2017: 1–10.

鄭蕭楨. 無人機(jī)航拍視頻序列意見征集[C]. AVS M4503, 大連, 中國(guó), 2016: 1.

ZHENG Xiaozhen. Opinion collection of aerial video[C]. AVS M4503, Dalian, China, 2016: 1.

ROSEWARNE C, SHARMAN K, and FLYNN D. Common test conditions and software reference configurations for HEVC range extensions[C]. JCTVC-P1006, San Jose, USA, 2014: 1–11.

BJ?NTEGAARD G. Calculation of average PSNR differences between RD-curves[C]. VCEG-M33, Austin, USA, 2001: 1–4.

BJ?NTEGAARD G. Improvements of the BD-PSNR model. ITU-T SG16 Q.6 document[C]. VCEG-AI11, Berlin, Germany, 2008: 1–2.

吳成家, 陳大鵬, 陳芳, 等. 一種計(jì)算綜合PSNR的方法[C]. AVS M3830, 杭州, 中國(guó), 2016: 1–11.

WU Chengjia, CHEN Dapeng, CHEN Fang, et al. A method for calculating integrated PSNR[C]. AVS M3830, Hangzhou, China, 2016: 1–11.

相關(guān)文章

施引文獻(xiàn)

資源附件(0)

訪問統(tǒng)計(jì)