Spatial Error Concealment Based on Edge Visual Clearness for Image/Video Communication
- PDF / 402,126 Bytes
- 10 Pages / 439.37 x 666.142 pts Page_size
- 49 Downloads / 183 Views
Spatial Error Concealment Based on Edge Visual Clearness for Image/Video Communication Ján Koloda · Victoria Sánchez · Antonio M. Peinado
Received: 24 February 2012 / Revised: 20 September 2012 / Published online: 12 October 2012 © Springer Science+Business Media New York 2012
Abstract In this paper, we propose a technique for concealing missing image/video blocks based on the concept of visual clearness of an edge. A scanning procedure based on the Hough transform allows us to find the relevant edges, and the visually clearest ones are employed in an interpolation based reconstruction. Specifically, several interpolations are combined according to a set of weights which allows the reconstruction of more complex textures. These weights are derived from the visual clearness associated to an edge and are unique for every pixel within the missing macroblock. The resulting algorithm is quite efficient, simple, and competitive in comparison with other state-of-the-art techniques. Keywords Error concealment · Directional interpolation · Block-coded image/video · Hough transform
1 Introduction Multimedia transmission applications are prone to suffer from deterioration of QoS. Due to strict real-time requirements, the retransmission of lost or severely damaged packages can be impossible. The block-based video coding standard H.264/AVC has introduced several error resilience mechanisms that draw on specific data organization tools such as network abstraction layer units (NALU), flexible macroblock orderJ. Koloda () · V. Sánchez · A.M. Peinado Department of Signal Theory, Networking and Communications and CITIC-UGR, Universidad de Granada, 18071 Granada, Spain e-mail: [email protected] V. Sánchez e-mail: [email protected] A.M. Peinado e-mail: [email protected]
816
Circuits Syst Signal Process (2013) 32:815–824
ing (FMO), or arbitrary slice ordering (ASO) [4, 17]. These tools allow the decoder to apply error concealment (EC) algorithms [8], in order to achieve an acceptable visual quality of the received stream. The EC algorithms benefit from the fact that video signals are highly correlated, spatially and/or temporally. This criterion is used to classify the EC algorithms into two groups: spatial EC (SEC), which utilizes only the information provided by the current frame, and temporal EC (TEC), which makes use of temporal information such as motion vectors (MV). This classification is nonexcluding, and combining temporal and spatial information leads to significant improvements [6]. However, the most extended block-based coding standards, such as H.264 and MPEG-4, use both intracoding (I-frames) and prediction (P/B-frames). Since the intracoded frames serve as a “firewall”, that separates visually different scenes or resets the prediction error, utilizing temporal information for their concealment could be risky. Therefore, SEC techniques are the most suitable choice to conceal the I-frames [17]. Moreover, these frames are employed as templates to predict several consecutive P/B-frames [4], so a poor reconstruction of one I-frame wo
Data Loading...
