An efficient video coding technique using a novel non-parametric background model
- Authors: Chakraborty, Subrata , Paul, Manoranjan , Murshed, Manzur , Ali, Mortuza
- Date: 2014
- Type: Text , Conference proceedings
- Relation: 2014 IEEE International Conference on Multimedia and Expo Workshops, ICMEW 2014; Chengdu; China; 14th-18th July 2014 p. 1-6
- Full Text:
- Reviewed:
- Description: Video coding technique with a background frame, extracted from mixture of Gaussian (MoG) based background modeling, provides better rate distortion performance by exploiting coding efficiency in uncovered background areas compared to the latest video coding standard. However, it suffers from high computation time, low coding efficiency for dynamic videos, and prior knowledge requirement of video content. In this paper, we present a novel adaptive weighted non-parametric (WNP) background modeling technique and successfully embed it into HEVC video coding standard. Being non-parametric (NP), the proposed technique naturally exhibits superior performance in dynamic background scenarios compared to MoG-based technique without a priori knowledge of video data distribution. In addition, the WNP technique significantly reduces noise-related drawbacks of existing NP techniques to provide better quality video coding with much lower computation time as demonstrated through extensive comparative studies against NP, MoG and HEVC techniques.
Undecoded coefficients recovery in distributed video coding by exploiting spatio-temporal correlation: a linear programming approach
- Authors: Ali, Mortuza , Murshed, Manzur
- Date: 2013
- Type: Text , Conference proceedings
- Relation: Proceedings of IEEE International Conference on Digital Image Computing: Techniques and Applications (DICTA 2013), Hobart, November 26-28th, 2013, p 1-7
- Full Text: false
- Reviewed:
- Description: Distributed video coding (DVC) aims at achieving low-complexity encoding in contrast to the existing video coding standards' high complexity encoding. According to the Wyner-Ziv theorem this can be achieved, under certain conditions, by independent encoding of the frames while resorting to joint decoding. However, the performance of a Wyner-Ziv coding scheme significantly depends on its knowledge about the spatio-temporal correlation of the video. Unfortunately, correlation statistics in a video widely varies both along the spatial and temporal directions. Therefore, we argue that in a feedback free transform domain DVC scheme the decoder will fail to recover all the transform coefficients with a nonzero probability. Thus, we suggest to integrate a recovery method with the decoder that aims at recovering the undecoded coefficients by exploiting the spatio-temporal correlation of the video. Besides, we extend and modify a recovery scheme, recently proposed in the context of images, for DVC so that it exploits both spatial and temporal correlations in recovering the undecoded coefficients. The essential idea of this scheme is to formulate the recovery problem as a linear optimization problem which can be solved efficiently using linear programming. Our simulation results demonstrated that the proposed scheme can significantly improve the PSNR and visual quality of the erroneous video frames produced by a DVC decoder.
Verifiable and privacy preserving electronic voting with untrusted machines
- Authors: Murshed, Manzur , Sabrina, Tishna , Iqbal, Anindya , Ali, Mortuza
- Date: 2013
- Type: Text , Conference proceedings
- Relation: Proceedings of the 12th IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom 2013) Melbourne, Vic, 16-18th July, 2013 p. 798-804
- Full Text: false
- Reviewed:
- Description: Designing a trustworthy voting system that uses electronic voting machines (EVMs) for efficiency and accuracy is a challenging task. It is difficult, if not impossible, to ensure the trustworthiness of EVMs that possess computation, storage, and communication capabilities. Thus an electronic voting system that does not assume trusted EVMs is clearly desirable. In this paper, we have proposed a k-anonymized electronic voting scheme that achieves this goal by assuming a hardware-controlled trusted random number generator external to the EVM. The proposed scheme relies on a k-anonymization technique to protect privacy and resort to joint de-anonymization of the votes for counting. Since the joint de-anonymization takes into account all the votes, it is difficult to manipulate an individual vote, even by the EVM, without being detected. Besides the anonymization technique, the proposed scheme relies on standard cryptographic hashing and the concept of floating receipt to provide end-to-end verifiability that prevents coercion or vote trading.
Abnormal event detection in unseen scenarios
- Authors: Haque, Mahfuzul , Murshed, Manzur
- Date: 2012
- Type: Text , Conference proceedings
- Relation: 2012 IEEE International Conference on Multimedia and Expo Workshops (ICMEW), Melbourne, 9-13th July, 2012. pg 1-6
- Full Text: false
- Reviewed:
- Description: Event detection in unseen scenarios is a challenging problem due to high variability of scene type, viewing direction, nature of scene entities, and environmental conditions. Existing event detection approaches mostly rely on context-specific tuning and training. Consequently, these techniques fail to achieve high scalability in a large surveillance network with hundreds of video feeds where scenario specific tuning/training is impossible. In this paper, we present a generic event detection approach where the extracted low-level features represent the global characteristics of the target scene instead of any context-specific information. From the temporal evolution of these context-invariant features over a timeframe, a fixed number of temporal features are extracted based on the periodicity of significant transition points and associated temporal orders. Finally, top-ranked temporal features are used to train binary classifier-based event models. In this approach, supervised training and exhaustive feature extraction are required only once while building the target event models. During real-time operation in unseen scenarios, event detection is performed based on the trained event models by extracting the required features only. The proposed event detection approach has been demonstrated for abnormal event detection in completely unseen public place scenarios from benchmark datasets without additional training and tuning. Furthermore, the proposed event detection approach has also outperformed recent optical flow based event detection technique.
Impact on vertical handoff decision algorithm by the network call admission control policy in heterogeneous wireless networks
- Authors: Sharna, Shusmita , Murshed, Manzur
- Date: 2012
- Type: Text , Conference proceedings
- Relation: 2012 IEEE 23rd International Symposium on Personal, Indoor and Mobile Radio Communications, Sydney, Sept. 9th-12th 2012, pp.893-898
- Full Text: false
- Reviewed:
- Description: Vertical handoff plays an important role to provide seamless connectivity for a mobile user in an overlapped multinetwork environment. On the other hand in order to maintain network stability, efficient management of available radio resource becomes crucial as network operators want high network utilization and maximum profit generation. For vertical handoff management, existing research works considered these user centric vertical handoff decision algorithm and network centric call admission control as two isolated decision mechanisms in heterogeneous wireless environment. In this paper, however, we propose a correlation between vertical handoff decisions and call admission control policies. We have developed a novel vertical handoff decision model using the Markov decision process based vertical handoff decision algorithm by refining the optimality criterion to factor in the probabilistic consequence of the call dropping rates so that mobile-centric vertical handoff decision algorithm and network-centric call admission control can work through a feedback mechanism to maximize respective objectives in synergy.