Giaime Ginesu

Detection of foreign bodies in food by thermal image processing

This paper deals with the problem of detection of foreign bodies in food. A new method for inspecting food samples is presented, using thermographic images to detect foreign bodies that are not detectable using conventional methods. At first, the basic background of thermography is given. Then, experiments to obtain well-contrasted thermographic images of different food and foreign bodies are discussed. The main part of the present paper introduces specific image processing methods that show a good recognition power of foreign bodies within food. Results achieved with a small set of test images are presented. The results are promising and the methods work even on some poorly contrasted images. To compare the different image processing and recognition methods, a quality index is defined. On the test images the success of the presented methods is shown and the difference in recognition results can be measured using the introduced quality index.

A multi-factors approach for image quality assessment based on a human visual system model

In this paper, a multi-factor full-reference image quality index is presented. The proposed visual quality metric is based on an effective Human Visual System model. Images are pre-processed in order to take into account luminance masking and contrast sensitivity effects. The proposed metric relies on the computation of three distortion factors: blockiness, edge errors and visual impairments, which take into account the typical artifacts introduced by several classes of coders. A pooling algorithm is used in order to obtain a single distortion index. Results show the effectiveness of the proposed approach and its consistency with subjective evaluations.

QoE assessment of multimedia video consumption on tablet devices

In this paper we present a study aimed at assessing the Quality of Experience in video streaming when the user is employing a tablet devices. The study has been conducted on a database containing subjective assessment scores relative to 216 streaming sessions of video sequences encoded with H.264/AVC and corrupted by typical wireless channel transmission errors. Four 20sec-long video sequences at Common Intermediate Format (CIF) spatial resolution have been corrupted with 54 combinations of key parameters for the reference application: bitrate, packet loss rate, playout delay and transmission interruption. Subjective evaluations have been collected in compliance with ITU-T Recommendation P.910 through singlestimulus Absolute Category Rating (ACR) from 40 subjects. The videos were reproduced on two different tablet devices: on an Apple iPad 2 and on a Samsung Galaxy Tab GT-P1000. Subjective observations include the perceived quality considering all aspects of the fruition chain, from the coding to the environment and device specific issues. The evaluation results also provided several remarks that can be helpful in designing systems and applications for multimedia contents played back on tablets. The results show a good correlation between subjective observations and some impairments so that a simple QoE index is proposed. A correlation study between subjective assessment and objective metrics is also provided.

JPEG2000-coded image error concealment exploiting convex sets projections

Transmission errors in JPEG2000 can be grouped into three main classes, depending on the affected area: LL, high frequencies at the lower decomposition levels, and high frequencies at the higher decomposition levels. The first type of errors are the most annoying but can be concealed exploiting the signal spatial correlation like in a number of techniques proposed in the past; the second are less annoying but more difficult to address; the latter are often imperceptible. In this paper, we address the problem of concealing the second class or errors when high bit-planes are damaged by proposing a new approach based on the theory of projections onto convex sets. Accordingly, the error effects are masked by iteratively applying two procedures: low-pass (LP) filtering in the spatial domain and restoration of the uncorrupted wavelet coefficients in the transform domain. It has been observed that a uniform LP filtering brought to some undesired side effects that negatively compensated the advantages. This problem has been overcome by applying an adaptive solution, which exploits an edge map to choose the optimal filter mask size. Simulation results demonstrated the efficiency of the proposed approach.

Lossy to lossless SPIHT-based volumetric image compression

A new coding scheme for lossy to lossless compression of volumetric data using a three-dimensional packet wavelet transform is presented. The need of a unitary transform via the integer lifting scheme and the performance comparison of different integer filter kernels is discussed. A state-of-the-art coder, set partitioning in hierarchical trees (SPIHT), is considered and modified in order to adapt its tree-structure to the 3D packet wavelet structure defined. The algorithm is extensively tested and shows very good performance, both for lossy and lossless coding.

Streaming video over wireless channels: Exploiting reduced-reference quality estimation at the user-side

We propose a source rate control scheme for streaming video sequences over wireless channels by resorting on a reduced-reference (RR) quality estimation approach. It works as follows: the server extracts important features of the original video, which are coded and sent through the channel along with the video sequence and then exploited at the decoder to compute the actual quality; the observed quality is analyzed to obtain information on the impact of the source rate at the given system configuration; at the receiver, decisions are taken on the optimal source rate to be applied next at the encoder to maximize the quality as perceived at the user-side. The rate is adjusted on a per-window basis to compensate low-throughput periods with high-throughput periods so as to avoid abrupt video quality changes, which can be caused by sudden variations in the channel throughput. The use of the RR quality estimation represents the main novelty of the proposed work. This has the advantage of allowing the rate control to optimize the user-perceived video quality after all the streaming system impairments have affected the signal, including actual channel errors, playback buffer starvation occurrences and error concealment. This approach is new in this context, since in the past proposals video models are used to predict the relationships of the quality with the coding rate, channel errors and starvation occurrences. Numerical simulations show how the proposed approach is able to achieve results similar to those obtained with model-based approaches, but with the significant benefit of not requiring any knowledge on the signal and channel characteristics.

Addressing un-interoperability issues in QoE models: Is a layered modelling effective?

Un-interoperability issue is herein defined as the unfeasibility (or at least difficulty) of integrating the results of different modelling activities to broaden the range of scenarios where the integrated model can be applied. Given the huge number of parameters to be considered when defining a Quality of Experience (QoE) model, many different QoE models have been proposed in the literature. Each involves different scenarios and parameters, and specifies how such features are combined. In order to solve un-interoperability issues, we propose a layered modelling approach, which follows the well-known model for interoperable systems in the networking field. Each layer of the model is devoted to a specific quality domain and the interfaces with the other domains are well-defined in order to further simplify the combination of models related to different domains. Such model could be the reference for a general-purpose QoE management system. Experimental results for validating the proposed model are provided.

Objective assessment of the WebP image coding algorithm

Research on multimedia data coding is allowing for better performance in terms of compression ratio, coding features, and robustness against transmission errors. While rate-distortion performance is being improved at a slower pace if compared to what we were used to up to a decade ago, remarkable advances are being made by adding complex features, such as fast adaptive transforms, lossy to lossless coding, compressed domain processing, etc. One of the latest codec which is expected to improve on the state of the art is the WebP algorithm released by Google. In this paper we provide an objective evaluation of WebP, by comparing it with the JPEG family algorithms. From the results it appears that the performance of the proposed codec is in line with that of the alternative methods, without achieving any major improvement and lacking several features.

Efficient scrambling of wavelet-based compressed images: a comparison between simple techniques for mobile applications

Image scrambling is a fundamental task for several applications, from secure digital content transmission to mutual visual authentication. This paper proposes and evaluates several simple techniques with the aim of being efficient in respect to wavelet compression algorithms and fit for mobile applications. Then, the proposed methods are meant to comply with the average structure of wavelet-based coders and require very low computational capacity. Comparative results are provided both for evaluating the scrambling efficacy and the computational complexity.

Wavelet Domain Scrambling for Image-Based Authentication

A simple image scrambling method for image-based authentication (IBA) is proposed and evaluated. The devised scheme may be embedded into any wavelet-based IBA framework to provide mutual authentication with challenge-response architecture. The severe constraints on security, data transmission capability and user friendliness are addressed by the complete system, especially in the wireless environment. Data and application scalability is provided through the JPEG2000 standard and JPIP framework