Edge-Guided Image Gap Interpolation Using Multi-Scale Transformation

Abstract

-This paper presents improvements in image gap restoration through incorporation of edge-based directional interpolation within multi-scale pyramid transforms. Two types of image edges are reconstructed; (a) the local edges or textures, inferred from the gradients of the neighbouring pixels and (b) the global edges between image objects or segments, inferred using Canny detector. Through a process of pyramid transformation and downsampling, the image is progressively transformed into a series of reduced size layers until at the pyramid apex the gap size is one sample. At each layer an edge 'skeleton' image is extracted for edge-guided interpolation. The process is then reversed; from the apex, at each layer, the missing samples are estimated (an iterative method is used in the last stage of up-sampling), up-sampled and combined with the available samples of the next layer. Discrete cosine transform and a family of discrete wavelet transforms are utilized as alternatives for pyramid construction. Evaluations over a range of images, in regular and random loss pattern, at loss rates of up to 40%, demonstrate that the proposed method improves PSNR by 1 to 5 dB compared to a range of best published works.Index Terms-Error concealment, multi-scale DCT/DWT pyramid, edge detection, image gap recovery, packet loss concealment. I. INTRODUCTIONmage gap restoration have a wide range of applications that includes in-painting of missing or damaged segments in still images or the replacement of image data packet lost in transmission. Further examples of applications and environments where image gap restoration can be usefully applied are enhancement of distorted biomedical signals [1], restoration of archived damaged images [2] and packet loss concealment over internet protocol (IP) networks [3].A main current application of image gap restoration is packet loss concealment. Packet loss errors may occur due to network congestions or due to signal loss in mobile devices. IP networks are best-effort environments [4,5] where the packet delivery is not guaranteed. The rapid growth in demand for relatively high bandwidth image/video streaming applications over IP networks motivates the need for packet loss recovery and concealment in order to provide more reliable network services and more acceptable user experience [6].There are three broad approaches for mitigating the loss of quality in received images due to packet loss: (a) automatic request for retransmission (ARQ) of the lost packets, (b) error control via forward error correction (FEC) methods and (c) error concealment (EC) methods. The first method retransmits a copy of the damaged/lost packet and results in an increase in bandwidth and delay proportional to error rate [5]. This method can be used on request for retransmission in networks where there is an interaction between sender and receiver. The second category of methods, FEC, employs error correction coding to recover lost pixels from the received information. This implies that the pixel values in successive...

Blood Oxygen Concentration and Physiological Data Changes During Motion While Wearing Face Masks

“…Appropriate computational methods of signal and image processing together with machine learning tools are then applied to extract the desired information. These computational tools include methods of signal analysis in the time, frequency, and scale domains [20], [21], methods of digital filtering to reject noise components [22], and computational intelligence methods for extracting signal features.…”

Computational Intelligence in Metric Analysis of the Skull in the Context of Maxillofacial Surgery

“…Different mathematical methods include the use of selected computational algorithms, data preprocessing in the time and functional domains [24], [25], image processing methods [26], and machine learning methods for facial shape recognition. General computational intelligence tools [27]- [29] and deep learning methods [30] can be applied to extract features, classify them, and any associated statistical processing.…”

An Improved Image Filtering Algorithm for Mixed Noise

“…The method uses the advantages of wavelet transform and uses an unsupervised dictionary learning algorithm to create a dictionary for noise reduction, which shows very competitive denoising performance. Reference [25] presented improvements in image gap restoration through the incorporation of edge-based directional interpolation within multi-scale pyramid transforms and reconstructed two types of image edges. Reference [26] shows how to design a complex wavelet with good characteristics based on the important characteristics of dual-tree complex wavelet transform and explains a series of applications of dual-tree complex wavelet in signal and image processing.…”