ITU AI and multimedia Seminar: exploring new areas and cross SDO synergy

This seminar was held in 2022 year 1 month 18 Japan , Geneva time 13 these 17 when 20 It will be held online .

Conference home page link ：https://www.itu.int/en/ITU-T/Workshops-and-Seminars/2022/0118/Pages/default.aspx

international telecommunications union （ITU） International Organization for Standardization （ISO） And the International Electrotechnical Commission （IEC）JTC1 SC29 Organized a meeting on “AI And multimedia ： Explore new areas and cross SDO Together ” Seminar , On 2022 year 1 month 18 Japan , Geneva time 13 these 17 when 20 It will be held online . The seminar will be held on 2022 year 1 month 17 solstice 28 It's held online on the 21st ITU-T The first 16 The study group meeting was held at the same time .

background

AI Enjoy a natural close relationship with multimedia . These two fields have been nurturing each other , Stimulate each other's development . The connection between the two is natural 、 Intrinsic and complementary . They are all based on the use of the most advanced mathematical tools , For example, modeling 、 Transformation and optimization techniques ： Multimedia as AI It provides a broad application field , and AI Enable multimedia by embracing revolutionary technologies and rich use cases and scenarios .

today ,AI And multimedia ITU-T SG16、ISO/IEC JTC1/SC29 And other organizations are studying a group of very strategic topics .

AI It is completely changing many technical fields in multimedia , Affect the video 、 Audio data coding 、 Express 、 utilize 、 Services and applications .AI The algorithm can be embedded into traditional multimedia data coding 、 Tools for processing and generating methods ; However , from AI A complete new paradigm driven by can also become a reality in the near future , And spawn new standards .

Multimedia data constitutes the majority of all Internet data today , Consumed by humans and machines .SC29 Of JPEG and MPEG The working group has been studying innovative technologies such as machine video coding , It is believed that these technologies will be deployed in SG16 is （ Or about to ） The development of a variety of digital multimedia applications and services .

stay 2021 year 12 month 7 Held in Dubai on the th is called “ITU CXO meeting ” At the high-level meeting of senior industry representatives , Yes AI And multimedia , The following conclusions are drawn , And include CXO In the communique of the meeting ：CxO emphasize , They support ITU Promote closer cooperation between communities , To make AI/ML Rich multimedia standards , admit ITU、ISO and IEC Leadership in this area .

Close involvement will AI And multimedia as a comprehensive discipline , Instead of two independent disciplines for technical research and standardization SDO Are encouraged to build a platform ：

• Information sharing 、 Exchange of views to promote the establishment of mutual understanding ;
• Coordination and synchronization ;
• For Industry 、 academic circle 、 Increased awareness of regulatory agencies, etc , To discover new use cases and scenarios ;
• Explore the way forward , Open up new fields ;
• Solve problems of common concern , Such as privacy 、 trust 、 Security 、 Regulatory matters, etc .

Besides , Some have good ability and history in developing multimedia standards SDO After the recent important milestone, we are investigating the future direction . This is it. ITU-T SG16 and ISO/IEC JTC1/SC29 The situation faced , They are 2020 In the middle of, we worked together to successfully develop universal video coding （VVC）.

Different SDO Coordination and synchronization between them means not only a perfect mutual understanding , To avoid repetition and overlap , And there should be a pattern , Through a reasonable division of labor , Make each SDO The best use of the power of . for example , come from JPEG and MPEG The underlying basic technical standards of can be ITU-T SG16 Some of the key questions （ Such as Q5/16、Q12/16、Q23/16、Q21/16 and Q28/16） The work of provides useful building blocks .

therefore ,ITU-T SG16 and ISO/IEC JTC1/SC29 This joint seminar on artificial intelligence and multimedia is being organized , Hope to promote the work to be carried out in the common field of artificial intelligence and multimedia .

Conference speech

The meeting began with ITU-T SG16 chairman Noah Luo and ISO/IEC JTC1/SC29 chairman Gary Sullivan Opening remarks .

http://mpvideo.qpic.cn/0b2ekqaaaaaaseai6k7ourqvavgdabkaaaaa.f10002.mp4?dis_k=3e026403da20b695ff5a06f34fc8a15f&dis_t=1645152302&vid=wxv_2238618402556952580&format_id=10002&support_redirect=0&mmversion=false

The first lecture was given by Masahito Kawamori It brings “AI And digital medicine （AI and digital health）”, The speaker is ITU-T SG16 Questions 26 and 28 Reported by .

http://mpvideo.qpic.cn/0b2enaaaaaaayiahs6pouvqva2gdabuaaaaa.f10002.mp4?dis_k=4108b11f84c5658de11fe40ccf3f1b46&dis_t=1645152302&vid=wxv_2238523456684507141&format_id=10002&support_redirect=0&mmversion=false

The second lecture was given by Yuan Zhang It brings “ Machine oriented video coding （Video coding for machines）”, The speaker is ITU-T SG16 WP3 The co chairman of ,ITU-T SG16 Question 12 Reported by ,ISO/IEC JTC1/SC29/ WG2 VCM activity The co chairman of .

http://mpvideo.qpic.cn/0b2eneaaaaaawaahsmhou5qva2odabuqaaaa.f10002.mp4?dis_k=b238f4a7d972822b0f04f8ebc6c91bca&dis_t=1645152302&vid=wxv_2238524245280768003&format_id=10002&support_redirect=0&mmversion=false

The third lecture was given by Touradj Ebrahimi It brings “AI And multimedia JPEG strategic （JPEG strategy on AI and Multimedia）”, The speaker is ISO/IEC JTC1/SC29/WG1 JPEG Convener .

http://mpvideo.qpic.cn/0bc3pyaacaaao4ahwq7os5qva7wdaf7aaaia.f10002.mp4?dis_k=cc5cbefb0244f4e2926121b143c29ad6&dis_t=1645152302&vid=wxv_2238525046040510472&format_id=10002&support_redirect=0&mmversion=false

The fourth lecture was given by Wang Yuntao It brings “AI Enabling multimedia standardization and the way forward （Aspects of AI-enabled multimedia standardization and way forward）”, The speaker is ITU-T SG16 Question 5 Reported by .

http://mpvideo.qpic.cn/0b2enyaaaaaam4ahv57ouvqva3wdabxaaaaa.f10002.mp4?dis_k=1141901886a9d6317446e7f675b3fd82&dis_t=1645152302&vid=wxv_2238525058170437633&format_id=10002&support_redirect=0&mmversion=false

The fifth lecture was given by Mei Yan It brings “ Digital person ： Use cases and techniques （Digital human: use cases and technologies）”, The speaker is F.DH-FM and F.DH-2D, ITU-T SG16 The editor of .（ Due to the limited number of videos , This speech and the next speech are put in a video ）

The sixth lecture was given by Weimin Zhang It brings “SG16 Medium AI Latest status of performance benchmark work （AI performance benchmarking work in SG16 in its latest status）”, The speaker is ITU-T F.748.11, ITU-T SG16 The editor of .（ This speech starts from 17:21 Start ）

http://mpvideo.qpic.cn/0bc3lqaacaaaemajempowfqvaxgdafoaaaia.f10002.mp4?dis_k=49229e9c76da31fb02d0282c8d12b199&dis_t=1645152302&vid=wxv_2238634396729950211&format_id=10002&support_redirect=0&mmversion=false

After a short tea break , The seventh lecture was given by Andrew Segall It brings “JVET An overview of video coding technology based on neural network in （Overview of technologies considered in JVET's neural network-based video coding exploration）”, The speaker is Neural Network Video Coding activity in JVET of ITU-T SG16 Question 6 and ISO/IEC JTC1/SC29/WG5 The co chairman of .

http://mpvideo.qpic.cn/0b2eluaaaaaanyahus7ou5qvaxodaboqaaaa.f10002.mp4?dis_k=6dd1151058f392da6112003d8eae4c59&dis_t=1645152302&vid=wxv_2238525757159587847&format_id=10002&support_redirect=0&mmversion=false

The eighth speech was given by Werner Bailer It brings “ Represented by neural network MPEG Compress （MPEG compression of neural network representations）”, The speaker is ISO/IEC JTC1/SC29/WG4 Neural Network Coding activity Chairman .

http://mpvideo.qpic.cn/0b2elyaaaaaanqahusxou5qvaxwdabpaaaaa.f10002.mp4?dis_k=2a27713b9068068fbf109363eaef6c91&dis_t=1645152302&vid=wxv_2238525743922364422&format_id=10002&support_redirect=0&mmversion=false

The ninth lecture was given by João Ascenso It brings “JPEG AI Current status and roadmap of （JPEG AI, its current status and its roadmap）”, The speaker is ISO/IEC JTC1/SC29/WG1 JPEG AI activity Chairman .（ Due to the limited number of videos , This speech and the next speech are put in a video ）

The tenth lecture was given by Stuart Perry It brings “JPEG Pleno Current status and roadmap of point cloud coding （JPEG Pleno Point Cloud Coding, its current status and its roadmap）”, The speaker is ISO/IEC JTC1/SC29/WG1 JPEG Pleno Point Cloud activity Chairman .（ This speech starts from 20:11 Start ）

http://mpvideo.qpic.cn/0b2eoaaaaaaaqaajcdposnqva4gdabyaaaaa.f10002.mp4?dis_k=d006f017e4fc6a108c92362b0c47a3c0&dis_t=1645152302&vid=wxv_2238627716713627653&format_id=10002&support_redirect=0&mmversion=false

The eleventh lecture was given by Elena Alshina It brings “JVET and JPEG The evaluation and complexity evaluation method of coding technology based on neural network （Methodologies for evaluation and complexity assessment of neural network-based coding technology in JVET and JPEG）”, The speaker is Neural Network Video Coding activity in JVET of ITU-T SG16 Question 6 and ISO/IEC JTC1/SC29/WG5 The co chairman of ,ISO/IEC JTC1/SC29/WG1 JPEG AI activity The co chairman of .

http://mpvideo.qpic.cn/0bc3n4aacaaaniai4ihosjqva36dafxqaaia.f10002.mp4?dis_k=8db293794abb959e8ccf15398645b184&dis_t=1645152302&vid=wxv_2238615179200774145&format_id=10002&support_redirect=0&mmversion=false

Attach conference speech PPT Download link ：https://pan.baidu.com/s/1qhU7zB5NtfGx2G1eoNI3Ow Extraction code ：0kb4