Multimedia Coding

Become proficient in the principles and techniques of coding and compression of multimedia content, acquiring the skills needed to understand, design, and evaluate media coding systems for images, video, and audio.

General educational objectives of the course in terms of expected learning outcomes.

1. Knowledge and understanding: the aim of the course is to provide students with the knowledge required to understand the theoretical and algorithmic principles underlying the coding and compression of multimedia content, with reference to entropy coding, lossy compression, predictive and perceptual coding techniques, as well as more recent compression paradigms based on machine learning techniques.
2. Applying knowledge and understanding: students will acquire the skills needed to implement, use, and evaluate techniques and systems for the coding and compression of images, video, and audio signals.
3. Making judgements: through classroom examples and experimental activities, students will be able to assess whether the coding and compression solutions they propose meet appropriate requirements in terms of quality, efficiency, and complexity.
4. Communication skills: students will acquire the necessary communication skills and appropriate technical terminology in the field of media coding and multimedia content compression.
5. Learning skills: the course aims to provide students with the theoretical and practical methodologies needed to independently address and solve new problems that may arise during professional activities. To this end, several topics will be discussed in class by actively involving students in the search for possible solutions to real-world problems.

Classroom lessons

Seminars

Should teaching be carried out in mixed mode or remotely, it may be necessary to introduce changes with respect to previous statements, in line with the programme planned and outlined in the syllabus.

Access to the teaching materials provided by the instructor is available on MS Teams, in the "Multimedia e Laboratorio" Team, code: qv1shk6.

All communications will take place through the official Telegram channel of the course, so students are requested to join it: https://t.me/+b_WjQqfJxqw4ODRk.

The course introduces the main principles, techniques, and paradigms for coding and compressing multimedia content, with particular emphasis on images, video, and audio.

Starting from a review of selected fundamental concepts in information theory, entropy coding, quantization, and transforms, the course will explore techniques for statistical data modeling, advanced entropy coding, lossy compression, rate-distortion optimization, predictive and perceptual coding, and residual coding. The main paradigms of traditional multimedia coding will also be discussed through selected case studies.

Part of the course will be devoted to machine learning-based compression techniques, with particular reference to latent representations, learned probabilistic models, and neural representations. Methods for the assessment of the quality and performance of coding systems will also be covered, including perceptual criteria and subjective evaluation.

Finally, selected emerging paradigms in media coding will be introduced, including perceptual, generative, and semantic approaches.

1. K. Sayood, Introduction to Data Compression, 5th ed., Morgan Kaufmann/Elsevier, 2017 - main reference textbook for classical principles and techniques.
2. Z.-N. Li, M. S. Drew, J. Liu, Fundamentals of Multimedia, 3rd ed., Springer, 2021 - complementary reference for image, video, and audio coding.
3. Y. Yang, S. Mandt, L. Theis, An Introduction to Neural Data Compression, 2023 - main reference for learned and generative compression.
4. Scientific papers and teaching material provided by the instructor during the course - for perceptual metrics, neural representations, semantic coding, and emerging topics.

In order to take the exam, in accordance with the regulations, it is MANDATORY to register on the Smart Edu portal and on any other platform, as MS Forms, required by the instructor to optimize logistics. For any technical issues regarding Smart Edu registration, please contact the Education and Student Services Office.

The examination will consist of an oral assessment and the evaluation of a project assigned to the student.

For the assignment of grades for individual assessments, the following criteria are typically followed:

Fail: The student has not acquired the basic concepts and is unable to complete the exercises.

• 18-23: The student demonstrates a minimal mastery of the fundamental concepts; their ability to present and connect content is modest, and they can solve simple exercises.
• 24-27: The student shows a good grasp of the course content; their ability to present and connect the content is good, and they solve exercises with few errors.
• 28-30 with honors: The student has acquired all course content and can present them comprehensively with a critical perspective; they solve exercises completely and without errors

The assessment of learning may also be conducted remotely if the conditions require it.

Students with disabilities and/or learning disorders (DSA) must contact the instructor and the CInAP representative at DMI well in advance of the exam date to inform them of their intention to take the exam with the appropriate compensatory measures.

1. Explain the role of probabilistic modeling in data compression and discuss how the knowledge or estimation of symbol distributions affects the efficiency of entropy coding. 
2. Describe the rate-distortion trade-off in a lossy compression system and discuss how it is addressed in both traditional codecs and modern learned compression systems. 
3. Explain the principles of predictive coding and residual coding, discussing why coding the prediction error can be more efficient than coding the original signal directly.