News

Netflix ingest and encoding pipeline is a cloud-based platform that generates video encodes for the Netflix streaming service. Due to the large throughput of the system, automated video quality assessment of the source videos and the generated encodes is essential in ensuring the quality of experience of Netflix subscribers. Owing to the diversity of video content, traditional video quality assessment methods are not able to meet the need of Netflix video pipeline. A joint research team between USC MCL and Netflix is assembled to tackle this difficult problem.

A scalable solution of video quality assessment method is proposed by Joe Yuchieh Lin (USC), Eddy Chihao Wu (USC), Dr. Ioannis Katsavounidis (Netflix), Dr. Zhi Li (Netflix), Dr. Anne Aaron (Netflix), and Prof. C.-C. Jay Kuo (USC). This method is called Ensemble-Learning-based Video Quality Assessment Index (EVQA). EVQA adopts a frame-based learning mechanism to address the limited training data problem and fuses multiple image quality assessment indices to generate the final video quality score. The superior performance of the proposed EVQA index is demonstrated by experimental results conducted on both LIVE and MCL-V video databases.

The discussion of Netflix’s pipeline system, current solutions and remaining challenges will appear in ICIP 2015. The work of EVQA is accepted by ICME 2015 workshop on Cloud-based Media.

Same with Zhenxia Zhou, Shanglin Yang survived the fierce interview last week with his outstanding performance in the test, as well as his solid background in academic research. Let us hear what he said about the internship in the coming summer.

1. Could you briefly introduce yourself? (Previous research/project experience, research interest and expertise)

Hi, my name is Shanglin Yang, an EE master student at USC admitted in Fall 2014. I earned the Bachelor Degree of Engineering from Zhejiang University, Hangzhou, China. Although most of my previous learning and research experiences were subjected to hardware in my undergraduate study, I obtained a chance to learn about signal processing and image processing in the “Freescale Cup Intelligent Car Racing” project.

After that, I got involved in the project “Research and Design of Micro Four-Rotor Aircraft” and finished my graduation thesis titled “Investigation and Design of Innovation and Networking Platform of Electric Machines”. During the process, the charm of the dynamic image processing, data modeling and programing intrigued me so much that I changed my major to signal processing in my graduate study at USC.

Currently, I am also interested in AR system and have worked on several related projects under Epstein Department of Industrial and Systems Engineering,

2. What was your first impression of USC and MCL?

USC has a great academic atmosphere for engineering study. She also has famous alumni in various fields, offering her students with significant resource and motivation to pursue higher goals. Challenges and opportunities coexist everywhere in this university.

I was impressed by the size of MCL at first, especially by the number of great projects finished in the past. It convinces me that the lab is highly productive and I am very pleasant to work with [...]

MCL research on Wi-Fi MAC design for its coexistence with Bluetooth was recently published by the IEEE Transactions on Computers (Vol 64, No. 2, 2015 February). This paper is entitled with “Coexistence Wi-Fi MAC Design for Mitigating Interference Caused by Collocated Bluetooth,” and co-authored by Dr. Chia-Chun Hsu, Professor David Wei of Fordham University, and Professor Jay Kuo.

Dr. Chia-Chun (Alex) Hsu, a former PhD alum of MCL, graduated from USC in 2007. He is currently an engineer at Mdiatek, Taiwan. Before joining Mediatek, he worked at HTC in Taiwan. His expertise is on the design of network protocols in Wi-Fi and LTE communication systems. Dr. David Wei is a Professor of Computer and Information Sciences at the Fordham University, Bronx, New York. Dr. Jay Kuo is the MCL Director.

The main contribution of this work is the development of a cross-layer adaptive packet fragmentation algorithm so as to reduce the effect of interference to the Bluetooth devices in the Wi-Fi network. It is an extension of an earlier paper published in the IEEE Wireless Communications and Networking Conference 2007 entitled with “A cognitive MAC protocol using statistical channel allocation for wireless ad-hoc networks.” This conference paper has many follow-up works as evidenced by its high citation number (around 200 times).

Congratulations to Hyunsuk Ko, a current MCL member, for getting an offer from the Electronics and Telecommunications Research Institute (ERTI).ETRI is one of the best known Korean research labs in the area of information and communications technologies with international reputation and visibility. Hyunsuk has accepted the offer and will join the Department of “Realistic Broadcasting Media Research” as a senior researcher. He will conduct research on 3D/UHDTV broadcasting technology, high-efficiency/scalable video coding etc. in the next several years.

Hyunsuk said, “Most companies are actively hiring good candidates nowadays. There are more openings for software engineers than any other positions because giant IT companies such as Google, Amazon, Facebook etc. are leading the market. On the other hand, if you have an expertise in big data analysis, machine learning and computer vision, the chance to get a decent offer increases significantly, which means that MCL members do have a bright future.”

Hyunsuk got offers from Korean companies such as Samsung and LG and Korean national research lab, ETRI. He had interviews with US companies such as Apple, Google, and Amazon as well. After pondering his future career between ETRI and these IT companies, Hyunsuk eventually decided to join ETRI to be an academic researcher rather than being an engineer. Besides, his research background fits the need of the ETRI team very well.

Looking backward on his PhD life, Hyunsuk said that his major growth came from the process of being trained as an independent researcher. Hyunsuk got his MS degree by working on H.264/AVC video coding only. In contrast, he conducted research on quite a few topics during his PhD years, such as 2D-to-3D conversion, HEVC-based 3D video, objective quality metric development for stereoscopic images, and other stereovision [...]

In March 2015, MCLab has a new visiting scholar, Prof. Songbin Li. Prof. Li is Professor at Institute of Acoustics, Chinese Academy of Science, China. We are glad to have an interview with him, talking about his previous research experiences and future expectations at USC.

1. Could you briefly introduce yourself and describe your previous research experience?

I come from the Institute of Acoustics (IOA) of the Chinese Academy of Sciences (CAS). Before joining IOA in 2012, I received my Ph.D. from the National Network New Media Engineering Research Center (3NMedia), Institute of Acoustics, Chinese Academy of Science in 2010 and then I worked as a post-doc in the Department of Electronic Engineering at Tsinghua University. My research focuses on multimedia signal processing especially on multimedia forensics and security, as well as steganography and steganalysis. I am also interested in creative technology related to network new media, such as computer animation and film special effects production technology.

2. What is your first impression of USC and MCLab?

USC is a world-famous university and the campus is very beautiful. USC is full of diversity, where students and faculties are from various countries and ethnic groups. As for MCLab, Prof. Kuo has, to a remarkable extent, a global reputation in multimedia technology area. Under his lead, the MCLab is vital and high-yielding of scientific papers. The students in MCLab are all very smart, enthusiastic and hard-working. I think they will make a difference in the future.

3. What is your future expectation and plan in MCLab? ​

I have several aims of my visit. First, I would like to join a certain subgroup related to steganography and steganalysis, aiming to make some contribution to MCLab. Second, I hope to continue the research [...]