Intelligent media computing technology and application for media convergence

摘要

Media Convergence is the merging of mass communication outlets—print, television, radio, and the Internet—along with portable and interactive technologies through various digital media platforms. As a new influential mainstream media, Media Convergence has now become a national strategy to integrate multiple media forms into one platform. Ideally, the intelligent media computing technology and application, including 5G, Augmented Reality/Visual Reality, Natural Language Processing, Computer Vision, Robotics, Big data, and Machine/Deep/Reinforcement/Transfer learning, should evolve into a knowledge base for purposes of delivering a dynamic experience and innovating media communication methods. However, how does one integrated media provide effective algorithm structures and tools that could merge, transform, and process various media forms, that is, cross-modal/multi-modal learning and representation? This question remains to be answered. This Special Issue covers innovative proposals or global views of well-established systems/approaches, and the approaches relying on computational or mathematical models. It consists of five articles which are selected for publication after multiple rounds of peer review and scrutiny. An overview of these articles is discussed in the following. The first article introduces different types of traceability models, considers the efficiency and storage security between different models, and then discusses the supervision of traceability information [1]. This article proposes a DAG blockchain consensus mechanism based on random verifiable function and witness mechanism. This article considers the related threats under the blockchain model, and finally analyses the principle of this model to resist witch attacks and random number attacks. The second article integrates recommendations and trust relationship into recommendations systems and proposes a new method based on dynamic bidirectional heuristic trust path search algorithm [2]. It studies the search problem of reliable trust path from the perspective of heuristic search. The proposed method can quickly find a trust path to help users make correct decisions in the interaction based on the predicted trust, which is realised by combining the trust mechanism with the intelligent search algorithm. This article also discusses further research about intelligent recommendation, so as to better adapt to the rapid development of media and the internet. The third article researches different types of feature selection methods, considering various sparse learning algorithms, followed by a novel fast classification method for fine grained emotion recognition tasks [3]. The authors confirmed that, using this faster dictionary learning algorithm, the convergence of the network model is accelerated, and a better balance between classification efficiency and classification quality is achieved. The article concludes discussing open problems that can trigger further research on the topic. The fourth article proposes a new music generation network based on transformers and guided by the music theory to produce high-quality music work [4]. In this study, the decoding block of the transformer is used to learn the internal information of single-track music, and cross-track transformers are used to learn the information amongst the tracks of different musical instruments. A reward network based on the music theory is proposed, which optimises the global and local loss objective functions while training and discriminating the network so that the reward network can provide a reliable adjustment method for the generation of the network. The method of combining the reward network and cross entropy loss is used to guide the training of the generator and produce high-quality music work. Compared with other multi-track music generation models, the experimental results verify the validity of the model. The fifth article focusses on the sensitivity protection in medical data p