Talking head generation (THG) has emerged as a transformative technology in computer vision, synthesizing realistic human faces synchronized with audio, image, text, or video inputs. This paper systematically reviews THG methodologies and frameworks, categorizing approaches into 2D-based, 3D-based, Neural Radiance Fields (NeRF)-based, diffusion-based, and other techniques. We explore the most effective approaches in THG, emphasizing training techniques that improve realism, identity preservation, and motion accuracy. THG has vast potential applications, including creating digital avatars, dubbing videos, enhancing virtual assistants, and improving video calls. However, challenges include needing large models, handling extreme head movements, maintaining language synchronicity, and ensuring smooth visuals persist. This review provides a comprehensive overview of current progress, identifies ongoing challenges, and suggests future research directions, including developing simpler and more adaptable models, enabling real-time processing on smaller devices, and creating ethical guidelines. By summarizing existing research and highlighting ongoing challenges, this overview offers valuable insights for anyone interested in future of talking head technology. For the complete survey, code, and curated resource list, visit our GitHub repository: https://github.com/VineetKumarRakesh/thg.

With the proliferation of multiple wireless technologies and smartphones with enhanced radio capabilities, it has become imperative to deliver services that can work seamlessly across heterogeneous networks. As smartphones advance in their capabilities, there is an increasing adoption of services with realtime delivery constraints (e.g. video telephony, streaming). In a heterogeneous network environment, latencies during handoff across networks becomes a critical impediment for delivery of real-time services. An important contributor to the hand-off latency is the authentication of a mobile node with the target domain. In this paper, we propose a proxy based approach to reduce the pre-authentication latency which is also designed to operate across domains with minimal trust relationship. Our approach targets latencies attributable to computation and over-the-air communication. Our solution, based on IEEE 802.21 Media Independent Handover and using EAP-TLS as the preferred authentication scheme, provides predictable performance guarantees under widely varying network conditions.

The increasing density and data rate of unlicensed band wireless devices in small office and home (SOHO) environments has led to significant inter- and intra-radio interference problems. Multiple competing standards such as the IEEE 802.11b/g, Bluetooth and ZigBee, all of which operate in the 2.4 GHz ISM band, can interfere with each other when used in typical indoor environments, potentially causing significant performance degradation. This paper presents detailed experimental results (using the ORBIT radio grid testbed) to quantify the effects of such interference in representative SOHO scenarios. In particular, different topologies, traffic loads and number of interfering devices are emulated to show the impact of multi-radio interference and to characterize each kind of interference. Further, a cross-layer, multi-radio interference diagnosis framework (called “spectrum MRI”) is described with the aim of isolating and classifying multi-radio interference problems using heuristic and model-based methods. A specific example of identifying interference problems which may affect an 802.11g video link is given to illustrate the proposed measurement and diagnosis framework.

The uniqueness of security vulnerabilities in ad hoc networks has given rise to the need for designing novel intrusion detection algorithms, different from those present in conventional networks. In this work, we propose an autonomous host-based intrusion detection system for detecting malicious sinking behavior. The proposed detection system maximizes the detection accuracy by using cross-layer features to define a routing behavior. For learning and adaptation to new attack scenarios and network environments, two machine learning techniques are utilized. Support Vector Machines (SVMs) and Fisher Discriminant Analysis (FDA) are used together to exploit the better accuracy of SVM and faster speed of FDA. Instead of using all cross-layer features, features from MAC layer are associated/correlated with features from other layers, thereby reducing the feature set without reducing the information content. Various experiments are conducted with varying network conditions and malicious node behavior. The effects of factors such as mobility, traffic density, and the packet drop ratios of the malicious nodes are analyzed. Experiments based on simulation show that the proposed cross-layer approach aided by a combination of SVM and FDA performs significantly better than other existing approaches.

With the increasing convergence of network, services, and devices, more number of end user devices are accessing digital media content which were hitherto accessible mainly from computers and television sets. This makes the digital media more prone to illegal copying and distribution. In this paper, we propose a robust digital video watermarking scheme for copyright protection using the principles of visual cryptography, scene change detection in videos, and conversion of each scene into an image by averaging the frames in the scene. In the proposed method, data is extracted and stored as one share of the watermark pattern instead of embedding data into the original video as in the conventional watermarking schemes. Since no data is embedded, the method is robust against watermark attacks aimed at detecting and destroying or distorting data embedded into videos. Simulation results show that the method is robust even in the case of 50% frame drop. All the watermarked scenes, not frames, need to be dropped for an effective attack against the watermark. Dropping of scenes makes the video less attractive for illegal copying and distribution.

Traditional digital rights management (DRM) systems are only two party systems, involving the owner and consumers. However, for scalability of business it is often necessary to involve additional levels of distributors and sub-distributors, who can promote and distribute the content in regions unknown to the owner. Thus, we propose an architecture for multiparty multilevel DRM system. The term 'multiparty' refers to involvement of many parties such as the owner, distributors, sub-distributors and consumers and the term 'multilevel' refers to multiple levels of distributors/sub-distributors. The architecture also supports the log files based violation detection, in case of violation of DRM system by any party. However, violation detection imposes a problem of preserving privacy of consumers. So, in the architecture, a provision is made to preserve their privacy.

In this paper, we propose a joint digital watermarking protocol for the multiparty multilevel DRM architecture using Garner's algorithm for the Chinese remainder theorem (CRT). Our protocol exploits the incremental nature of the computation of CRT by the Garner's algorithm. The proposed joint watermarking protocol embeds a single watermark signal into the content while taking care of the various security concerns such as proof of involvement in the distribution chain, nonrepudiation of the involvement and protection against false framing of the different parties involved. Further, in the event of finding an illegal copy of the content, the identities of all the parties involved in that content distribution chain can be traced back by extracting the watermark information.

For scalability of business, multiparty multilevel digital rights management (DRM) architecture, where a multimedia content is delivered by an owner to a consumer through several levels of distributors has been suggested as an alternative to the traditional two party (buyer-seller) DRM architecture.

In ad hoc networks, the vulnerability of nodes to routing attacks is a serious concern. In this work we propose a cross-layer based routing attack detection system for ad hoc networks. Previous work that uses mostly audit trails collected from the routing protocol suffers from inadequacy of features to construct a reliable model for detecting anomalous routing behavior. On the other hand, use of linear detectors lead to very high false positives and false negatives because of the inherent non-linear nature of the feature space. In this work, we address these issues by collating features from multiple protocols at different layers and using a non-linear detector based on support vector machine (SVM). The consequent problem of computational expense of the detection process is addressed by a combination of novel data reduction techniques. Simulation results show that the performance of the proposed CRADS is far superior than conventional protocol-specific detection systems.

A new time division multiple access (TDMA)-based MAC protocol using directional antennas (RTDMA-DA) is presented for wireless mesh networks. RTDMA-DA is fully- distributed, with the nodes needing to reserve time slots in order to transmit packets. To ensure fairness, it makes use of a novel approach to give nodes within their local 1-hop neighborhoods alternating priorities to reserve slots. In addition, RTDMA-DA allows such local reservations to take place throughout the network, making the protocol highly scalable. In the case of nodes with varying bandwidth requirements, dynamic slot reservations can take place, thus increasing the overall bandwidth utilization. The results of the simulations show that RTDMA-DA achieves a saturation throughput of up to 11 times more than 802.11 and is highly scalable.

In a CSMA/CA based multihop wireless network, excessive interference at a receiver or a potential forwarding node causes severe blocking and reduction in throughput. The unbalanced interference forces the node to consume more time receiving packets rather than sending them, resulting in dropped packets due to buffer overflow. We discuss a novel flow control framework at the MAC layer for regulating the transmission and improving the overall throughput of multihop wireless networks based on CSMA/CA protocol. The framework prevents congestion, reduces packet loss and is attractive because per-flow information at each node is kept to a minimum. The techniques used to improve throughput include a hop-by-hop, hybrid rate and window based flow control scheme that paces the transmission of frames such that competition between frames originating from the same flow is reduced. Simulations show more than five fold improvement in throughput and energy cost over the existing scheme without flow control in some cases

In this paper, we propose a multi-rate CSMA/CA-based MAC protocol, which we refer to as EMR, to enable fast forwarding of frames in a more efficient manner. We use the base-rate RTS and CTS frames to reserve a long transmission path and utilize intermediate nodes to relay the data frames using shorter transmission distance but with more efficient transmission rates. Results show that our EMR scheme can outperform the traditional multi-rate forwarding scheme by 18% and 49.4% in a low-density and high-density ad hoc network, respectively.

In this paper, we present a novel reversible watermarking scheme for image authentication for JPEG/JPEG-2000 coded images. Since the watermarking scheme is reversible, the exact original image can be recovered from the watermarked image. The watermarking scheme makes use of finite state machine principles. The proposed scheme is asymmetric as the watermark extraction key is different than its embedding key. The algorithm is implemented and tested for its visual quality, compression overhead, execution time overhead and payload capacity. It is found that the algorithm has high visual quality, high payload capacity, low compression overhead and low execution time overhead.

We present a novel geometric approach to the popular post nonlinear (PNL) BSS problem. A PNL mixing system includes two stages: a linear mixing followed by a nonlinear transformation. In our method, the process to linearize the nonlinear observed signals, the most critical task in PNL model, is carried out by a geometric transformation. The basic idea is that in a multi-dimensional space, a PNL mixture is represented by a nonlinear surface while a linear mixture is represented by a plane. Thus, by transforming a PNL's representing nonlinear surface to a plane, the PNL mixture can be linearized. The hidden sources are then estimated from linearized signals by a linear BSS algorithm. Experiments show promising performance of our approach.

A computational efficient artificial neural network for adaptive channel equalization in a digital communication system with 4-QAM signal constellation is purposed. We proposed a single layer Chebyshev neural network (ChNN) by expanding the input pattern by Chebyshev polynomials. Performance comparison was carried out through extensive computer simulations with two other neural networks: an MLP and a functional link ANN together with a linear LIMS-based equalizer. It is shown that the ChNN provides satisfactory results in terms of convergence rate, MSE floor and BER over a wide range of EVR, SNR and nonlinear conditions with substantial reduction in the computational complexity.

Usually the targets in forward looking infra-red imagery are dim, slowly moving, and buried under clutter and noise. Detecting and tracking of such targets is a challenging task. Although artificial neural networks (ANNs) have been used to solve this problem, they need a lot of training time. In order to reduce the training time, we propose principal component analysis as a dimension reduction technique. We used an MLP with LM learning algorithm and a RBF neural network (RBFNN) with K-means algorithm to cluster the data. Both the ANNs are used in a neural adaptive line enhancer (NALE) configuration. Extensive computer simulations showed the combination of PCA and ANNs gives satisfactory results with significant reduction in training time.

We propose two approaches for microarray gene expression analysis and gene selection using neural networks. Using these approaches, only those genes which help sample classification are selected from the original set of genes, and the redundant genes expression patterns involved in the huge microarray matrix are eliminated so that dimensionality of the matrix is reduced from a few thousands to a much smaller number. An unsupervised SOM based technique and another supervised single layer perceptron based technique have been utilized for this purpose. Performance of these two approaches is compared in terms of accuracy, implementation and execution time.

Broadcasting is one of the essential communication models of MANET. Many MANET multicast routing protocols rely heavily upon MAC layer's broadcast support. However, the broadcast mechanism of the standard IEEE 802.11 cannot provide reliable broadcasting service, and therefore, degrades the multicasting performance. In this paper, we propose an extension to the IEEE 802.11 broadcast mechanism, called round-robin acknowledge and retransmit (RRAR), to improve the reliability. Different from other mechanisms, we employ a simple and effective DATA/BrACK (broadcast acknowledgment) mechanism. Receiving the BrACK, the DATA sender calculates and records the indices of the frames that are lost and retransmits them. The proposed scheme provides a highly reliable broadcast service to the routing layer, and the performance of multicast routing protocols can be improved significantly.

Anomaly-based intrusion detection (AID) techniques are useful for detecting novel intrusions without known signatures. However, AID techniques suffer from higher false alarm rate compared to signature-based intrusion detection techniques. In this paper, the concept of intrusion context identification is introduced to address the problem. The identification of the intrusion context can help to significantly enhance the detection rate and lower the false alarm rate of AID techniques. To evaluate the effectiveness of the concept, a simple but representative scheme for intrusion context identification is proposed, in which the anomalies in the intrusive datasets are visualized first, and then the intrusion contexts are identified from the visualized anomalies. The experimental results show that using the scheme, the intrusion contexts can be visualized and extracted from the audit trails correctly. In addition, as an application of the visualized anomalies, an implicit design drawback in t-stide is found after careful analysis. Finally, based on the identified intrusion context and the efficiency comparison, several findings are made which can offer useful insights and benefit future research on AID techniques.

Image flicker, the unnatural temporal fluctuation in perceived image intensity, is an annoying artifact existing in many old film sequences and its removal is an integral part in any old film restoration process. In this paper, we propose an improved flicker removal method based on flicker parameters estimation with the aide of motion vectors compensation. We tested the proposed method on sequences with artificially added global and local flicker, as well as on natural real flicker film material. Test results obtained showed the proposed method produced much better flicker correction quality than an established flicker removal method known as the "Roosmalen method".

We define and analyze a special case of exposed node problem that causes severe throughput degradation, which we term as the "critically exposed node problem." In a simple chain setup, we show that the reduction of throughput due to this problem is up to a factor of 14 compared to the ideal condition. There are practically very few suggestions in the literature to tackle the exposed node problem and none address the critically exposed node problem. We show by simulation how the critically exposed node problem affects capacity. We then explore a cluster-based channel assignment approach that can be used along with a multi-channel MAC protocol, called primary channel assignment based MAC (PCAM), to solve the problem. Using theoretical analysis, we work out the minimal number of channels required using an ideal, overlapped-clustering scheme and the multi-channel MAC protocol. Finally, we propose a distributed sub-optimal heuristics-based channel assignment protocol to address this issue.

Broadcasting is one of the essential communication models of MANETs. Many MANET multicast routing protocols rely heavily upon MAC layer's broadcast support. However, the broadcast mechanism of the standard IEEE 802.11 cannot provide reliable broadcasting service. In this paper, we improve the IEEE 802.11 broadcast mechanism's reliability by introducing the adaptive round-robin acknowledge and retransmit (ARAR) scheme. Different from the other mechanisms, ARAR is able to adapt to the traffic load by adjusting the length of a frame transmitted. For short frames, a simple and effective DATA/BrACK (broadcast acknowledgment) scenario is applied; for long frames, RTS/CTS/DATA/BrACK exchange is utilized. Receiving the BrACK, the DATA sender calculates and records the indices of the frames that need retransmission and performs the retransmission when it is possible. The proposed scheme provides a highly reliable broadcast service to the routing layer, and the performance of multicast routing protocols can be improved significantly. This is shown through simulations, in which we have compared ARAR to the standard IEEE 802.11 approach using ODMRP [Y. Yi et al., Nov. 2002]. Our future work attempts to compare ARAR with some existing works.

Some wireless MAC protocols such as IEEE 802.11 were not designed to capitalize on the available multiple channels in an efficient manner. In this paper, we propose a novel multi-channel medium access control protocol for wireless networks which we term us "primary channel assignment based MAC". The protocol is based on the use of primary channel assignment, which is used by other nodes to find the corresponding node. We also introduce a method to reduce the hidden terminal problem encountered in multi-channel designs by proposing a modified transmission range threshold. The scheme eliminates the need for a separate dedicated control channel or time slot that is usually prone to saturation when the traffic increases. The proposed scheme also addresses the broadcast messages problem in multi-channel design. The scheme is based on IEEE 802.11 DCF protocol. The performance of the proposed protocol is compared to two other multi-channel schemes. In an ad-hoc network, the protocol outperforms the other protocols by up to 140%.

Information and infrastructure security is a serious issue of global concern. As the last line of defense for security infrastructure, intrusion detection techniques are paid more and more attention. In this paper, one anomaly-based intrusion detection technique (ScanAID: Statistical ChAracteristics of N-grams for Anomaly-based Intrusion Detection) is proposed to detect intrusive behaviors in a computer system. The statistical properties in sequences of system calls are abstracted to model the normal behaviors of a privileged process, in which the model is characterized by a vector of anomaly values of N-grams. With a reasonable definition of efficiency parameter, the length of an N-gram and the size of the training dataset are optimized to get an efficient and compact model. Then, with the optimal modeling parameters, the flexibility and efficiency of the model are evaluated by the ROC curves. Our experimental results show that the proposed statistical anomaly detection technique is promising and deserves further research (such as applying it to network environments).

This paper presents new MRF (Markov random field) models in a bidirectional Bayesian framework for accurate motion and occlusion field estimation. With careful selection of the five free parameters required by the models, good experimental results have been obtained. The resultant computational speed is also 5.5 times faster compared with the conventional "iterated conditional mode" relaxation using the proposed fast bidirectional relaxation.

Association rule mining is a well-researched area where many algorithms have been proposed to improve the speed of mining. In this paper, we propose an innovative algorithm called Rapid Association Rule Mining (RARM) to once again break this speed barrier. It uses a versatile tree structure known as the Support-Ordered Trie Itemset (SOTrieIT) structure to hold pre-processed transactional data. This allows RARM to generate large 1-itemsets and 2-itemsets quickly without scanning the database and without candidate 2-itemset generation. It achieves significant speed-ups because the main bottleneck in association rule mining using the Apriori property is the generation of candidate 2-itemsets. RARM has been compared with the classical mining algorithm Apriori and it is found that it outperforms Apriori by up to two orders of magnitude (100 times), much more than what recent mining algorithms are able to achieve.

With the rapid proliferation of mobile devices and services, efficient support for mobile transactions has become extremely important. As locking based pessimistic approaches to concurrency control have been known to be not very suitable for the mobile environment, many optimistic concurrency control schemes have been proposed to date, of which the broadcast certification based approach is of special interest because of its simplicity and its efficient use of the downlink bandwidth. The authors propose modifications to the broadcast certification based protocol and show, using simulation results, how useful tradeoffs can be obtained between client side and server side validation, resulting in optimal performance.

At present, there are no association rule mining algorithms that are suitable for use in electronic commerce because they do not consider that new products are introduced and old ones are retired frequently and they assume that support thresholds do not change. In this paper a new algorithm called Fast Online Dynamic Association Rule Mining (FOLDARM) is introduced for mining in electronic commerce. It uses a novel tree structure known as a Support-Ordered Trie Itemset (SOTrieIT) structure to hold pre-processed transactional data. It allows FOLDARM to generate large 1-itemsets and 2-itemsets quickly without scanning the database. In addition, the SOTrieIT structure can be easily and quickly updated when transactions are added or removed. It also stores data that is independent of the support threshold and thus can be used for mining with varying support thresholds without any degradation in performance. Experiments have shown that FOLDARM outperforms Apriori, a classic mining algorithm, by up to two orders of magnitude (100 times).

It is well known that accurate dense motion field can improve the video coding efficiency. This paper presents a novel Markov random field (MRF) model that estimates both the dense motion and uncovered background fields in image sequences, and the application of these estimates in H.263-based video coding framework.

This paper proposes a new Markov random field (MRF) model for the detection of occluded regions in image sequences. Motion vectors are not defined in an occluded region, thus the regions with high motion compensated prediction error are commonly regarded as occluded regions. However, badly motion compensated pixels will also appear as occluded pixels, making it difficult to distinguish the true occluded pixels from the poorly motion compensated regions. The proposed MRF model addresses this problem by incorporating motion information into the occlusion model. This is derived from the observation that occlusion occurs when objects move. It is found that an accurate occlusion region can be detected and better motion estimation can be made with the new model using iterated conditional modes (ICM).

We present a new class of interconnection topologies called the Linear Recursive Networks (LRNs) and examine their possible applications in distributed systems. Each LRN is characterized by a recursive pattern of interconnection which can be specified by simple parameters. Basic properties such as node degree, diameter, and the performance of routing algorithms for all LRNs are then collectively analyzed in terms of these parameters. By choosing appropriate values for the parameters, our results can assist a network designer in selecting a topology with required routing performance and cost of interconnection. A subclass of LRNs, called Congruent LRNs (CLRNs), is also identified here and shown to possess desirable properties for more tightly coupled systems. It is shown that the CLRNs include existing networks such as hypercube and generalized Fibonacci cubes. These results suggest that the linear recursive networks potentially have applications in interconnecting distributed systems.

We present a parallel algorithm for RNS to binary conversion, which is computationally efficient and requires moderate amount of storage. The algorithm is based on a graphical interpretation of the residue numbers, and imposes no restriction on the size or choice of moduli set. The parallel time complexity of the algorithm is /spl Theta/ upper bound [(log k)], where k is the size of the moduli set.>

We present a new class of interconnection topologies called the Linear Recursive Networks. It is shown that each member of this fairly large class of network topologies possesses a recursively-decomposable structure. The path and connectivity properties of the entire class of networks are analyzed and found to have useful features. It is also demonstrated that Point-To-Point or Broadcasting/Gathering operations can be performed efficiently on all Linear Recursive Networks. The preliminary results suggest that the Linear Recursive Networks potentially have many applications.