G.Y. Hong

An intelligent video categorization engine

Purpose – We describe an intelligent video categorization engine (IVCE) that uses the learning capability of artificial neural networks (ANNs) to classify suitably preprocessed video segments into a predefined number of semantically meaningful events (categories).Design/methodology/approach – We provide a survey of existing techniques that have been proposed, either directly or indirectly, towards achieving intelligent video categorization. We also compare the performance of two popular ANNs: Kohonens self‐organizing map (SOM) and fuzzy adaptive resonance theory (Fuzzy ART). In particular, the ANNs are trained offline to form the necessary knowledge base prior to online categorization.Findings – Experimental results show that accurate categorization can be achieved near instantaneously.Research limitations – The main limitation of this research is the need for a finite set of predefined categories. Further research should focus on generalization of such techniques.Originality/value – Machine understandin...

Interoperability in a Wireless Home Networking System for Healthcare Monitoring

Versatile wireless home networks provide a convenient means for keeping track of health with appropriate biosensors. It utilizes existing network for A/V and computing applications is presented in this paper. The system provides reliable communication channels for a number of consumer applications while optimal allocation of resources is provided with an interoperability specification for an IEEE 802.11 based network connecting to an external wireless telemedicine system. Each device is identified with an RFID tag and the system allocates necessary network resources with adequate QoS assurance for a given application.

Error localization for robust video transmission

The convergence of Internet, multimedia and mobile applications has led to an increased demand for efficient and reliable video data transmission over heterogeneous networks. Due to their coding efficiency, variable-length codes (VLC) are usually employed in the entropy coding stage of video compression standards. However, error propagation is a major problem associated with VLC. We propose the use of a class of self-synchronizing VLC (SSVLC) to achieve the dual goal of optimal coding efficiency and optimal error localization. Performance evaluation has confirmed that the use of SSVLC provides better performance than standard VLC techniques.

QoS control for Internet delivery of video data

Although the emerging Internet Protocol version 6 (IPv6) includes provision for quality-of-service (QoS) control capabilities, there is much uncertainty surrounding its widespread acceptance. The Internet is becoming an increasingly popular communication medium due to its wide geographic coverage and continuous availability. However, it is also a harsh environment for transmission of high-bandwidth video data due to severe and time-varying network constraints. Since the best-effort nature of the current Internet does not offer any QoS guarantees, it is necessary to provide sender/receiver-based QoS control to provide adequate video play-out quality at the receiver. In this paper, we investigate current QoS control techniques and propose an integrated approach to this problem.

Constrained error propagation for efficient image transmission over noisy channels

Multimedia applications increasingly require efficient transmission of still and moving images over wireless channels. However, wireless channels originally designed for voice transmission impose severe bandwidth constraints on image data. In addition, data transmitted over wireless channels are subject to degradation due to such phenomena as burst noise and fading that cause errors. The JPEG standard has been developed for image data compression necessary for efficient transmission. The emerging JPEG-2000 standard addresses the transmission error problem by including provisions for error-resilience tools. These include resynchronization, data recovery and error concealment. In this paper, we survey current error resilience techniques for image data, and focus on the synchronization aspect of error resilience. In particular, we propose an application of self-synchronizing variable length codes (VLC) to achieve the dual goal of efficient source coding and constrained error propagation to localize the effects of data corruption. This makes subsequent data recovery and error concealment more effective. Performance analysis of the proposed technique is also presented.

Adaptive QoS control of multimedia transmission over band-limited networks

Consumers increasingly demand high quality of service (QoS) for multimedia applications. Application-layer QoS control mechanisms that do not make specific references to coding semantics tend to provide the most robust solutions for delivery of multimedia data over band-limited networks. We propose an adaptive and robust QoS algorithm for band-limited networks that integrates explicit receiver-sender rate control, rate shaping with multiple quantization parameters and adaptive forward error correction (FEC). Experimental results indicate that an average improvement of 0.5 dB PSNR can be achieved compared to other methods.

Adaptive single carrier modulation scheme for high-speed mobile computing

We present an adaptive modulation scheme optimized for networking devices with high mobility, intended to improve power efficiency and bandwidth usage for various operating environments. We compare a number of deployment options and found a marginal improvement in power consumption and spectral utilization efficiency compared to fixed modulation.

Differential detection of single carrier signals for wearable computers

Wearable computers operating in an outdoor environment may experience uncontrollable changes of operating environments. Cost-effective solutions with fast synchronization properties make differential detectors particularly suitable for wearable computers. Our simulation results show that the proposed detection mechanism offers an adequate level of BER performance and gain for the receiver at a relatively low cost with low power consumption due to its simple receiver structure complexity.

Cost driven six sigma analysis for CE software development

Increased software demands from consumers lead CE (consumer electronics) companies towards a shift from hardware-oriented to a more software-oriented industry. Many CE companies spend a tremendous amount of money on software development and post release defect fixing. We propose a modification of the traditional six sigma approach to fully analyze and measure software development cost.

A QAM scheme for fast wireless Internet and multimedia services

Wireless wideband communication systems provide high data transmission rates suitable for fast Internet access. This paper presents a modulation scheme for Local Multipoint Distribution Service (LMDS), which improves coverage under rainy weather conditions. Following a comparison of a number of modulation schemes with a 26 GHz carrier, a 16-QAM system is found to offer good compromise between coverage and receiver structure complexity. The effects of rain attenuation increase with carrier frequency. System performance varies significantly depending on whether the system is operated on a dry day or under heavy rain.