News

MCL lab is proud to release a new dataset about compressed images. It consists of 50 source images with resolution 1920×1080 and 100 JPEG-coded images for each source image. More than 150 volunteers participated in the subjective test. Each individual set of compressed images was evaluated by 30 subjects in a controlled environment.

This dataset was proposed to challenge the traditional approaches to measure the quality of compressed image/video. Based on the characteristics of the Human Visual System (HVS), a Just Noticeable Difference (JND) framework was proposed to investigate the limitation of HVS on compressed images. It means to boost large-scale statistical study on human-perceived image quality as well as the future development of perceptual-based image/video coding standards.

Sachin Chachada, a MCL member, has passed his defense on Jan 14, 2016. Congratulations!
His dissertation title is “Classification and Retrieval of Environmental Sounds”. Speech and music signals have been extensively studied for several decades. Need for Environmental Sound Recognition (ESR) system has picked up the pace in recent years. Environmental sounds are quotidian sounds, both natural and artificial, i.e. sounds one encounters in daily life other than speech and music. In this thesis, Narrow Band Time Frequency features are proposed which characterize a signal using TF representation of its band limited filtered signal. In order to further improve the performance, use of a novel multi-classifier approach, Para-Boost (PB) model, is proposed. It takes the advantages of all the features and improves the overall performance of ESR system. Finally, considering the exponential growing environmental sound data on the Internet, the thesis tackles the problem of a good content based retrieval system. A two stage content based environmental sound retrieval system is proposed. This query-by-example retrieval system assumes that the database is partially labeled. In Stage I, a broad categorization of environmental sounds based on their signal-characteristic is done. For each category, a classifier is trained to predict labels for unlabeled data in the database and also narrow search range for a query by assigning it multiple, yet limited, class labels. In Stage II, a novel feature and a scoring scheme to do local matching and ranking is proposed. An audio signal is first segmented in an unsupervised manner using Mean Shift algorithm and each segment is represented by its point of convergence in the feature space. The audio signal is finally represented by its energy distribution over each segment thereby capturing the temporal variations [...]

Zhi Tong graduated as a master student from USC in May 2013. He worked as a research assistant at MCL from May 2013 to May 2015. He decided to go back to his home country and he recently got an offer from Changjiang Securities, one of the top 10 securities companies in China. Congratulations to Zhi! We got new year greetings from Zhi to all group members. We had an interview with him.

Q: Can you briefly introduce your current work, e.g., your responsibility, position, etc.?
A: My current role is a product manager at Changjiang Securities. My responsibility is designing transactions and management systems for prime brokers (e.g. private equity and trust companies), capital management and proprietary trading.

Q: What makes you stand out among other candidates applying for the same position?
A: First, good manners. It can give interviewee a good impression. Second, out-going characteristic. It shows you can communicate with colleagues effectively and suit for teamwork. Third and most importantly, clear logic, the ability to analyze problems and define problems.

Q: How’s the learning experience in USC and MCL differentiate you from your colleagues in your company?
A: I think when we are doing research with professor Kuo, we learn how to analyze problems and define problems. Such abilities are crucial and desired no matter what field you are in or what job you have. Second, good math and knowledge of machine learning. Recently, my boss is interested in using machine learning techniques to predict certain performances of stocks. Good thing for me is that only I can do it in my team.

Q: Can you share your job hunting experience in China ? Do you have any suggestions for group members who want to [...]

MCL Director, Professor C.-C. Jay Kuo, has been traveling in Asia and attended a couple of events in the last three weeks. He visited 5 Asian cities, i.e., Singapore, Chengdu, Beijing, Taipei and Hong Kong, in his trip. His activities include the delivery of three speeches: “Reflection on Image/Video Coding: Where Do We Go from Here?”, “Deep Learning: Hype or Hope?” and “Multimodal Image Retrieval in various conferences, campuses and company sites.

In particular, he attended the 10th International Conferences on Information, Communications and Signal Processing (ICICS) held in Singapore (Dec. 2-4) and the 7th Annual Conference of the Asia-Pacific Signal and Information Processing Association (APSIPA) held in Hong Kong (Dec. 16-19). His speeches were very well received.

One of his most enjoyable events was the re-union with MCL alumni in Taipei on Dec. 12.

Thanksgiving luncheon has been a tradition of MCL for over twenty years and this year is no exception. This Thursday (11/25/2015), Prof. Kuo held the Thanksgiving luncheon for all MCL members at China Great Buffet in El Monte. More than twenty people including Prof Kuo’s family and other guests invited by MCL members attended this luncheon. We enjoyed the tasty food there including sushi, sashimi, seafood, Chinese food and dessert. This was also a great opportunity for MCL members to get to know each other. We chatted casually and most topics were about Thanksgiving plans.

Although we could not spend Thanksgiving with our family, we really felt quite warm when celebrating Thanksgiving together with lab fellows.