A statistical model-based diagnostic scheme for cost-effective determination of freshness in fish industry

Dimogianopoulos, Dimitrios; Grigorakis, Kriton

doi:10.1109/med.2011.5982993

Cited by 2 publications

(20 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As explained in Dimogianopoulos and Grigorakis (), the function g ( ζ , ω n ) should use selected dominant pairs of continuous time poles, in terms of damping factor ζ and natural frequency ω n , for a simple reason: because the dominant poles account for the major part of the system dynamics, they are the most evident elements to be examined if potential changes in system dynamics should be monitored.…”

Section: Resultsmentioning

confidence: 99%

“…The procedure introduced in Dimogianopoulos and Grigorakis () for selecting the dominant pole pairs to be investigated involved plotting the initial three or four most dominant pole pairs on the pole zero plane and looking for visual differentiations between dominant pairs from fresh fish samples and their counterparts from less fresh ones. Figure presents the four most dominant continuous time pole pairs from the identified AR model with n y = 66.…”

Section: Resultsmentioning

confidence: 99%

“…The fish were tested following a procedure similar to that in Grigorakis and Dimogianopoulos (). Nonetheless, the technical solutions used for testing were inspired by those in Dimogianopoulos and Grigorakis (), as follows: a simple (and robust to overloading) spring‐based balance integrating an electropneumatic transducer was used. A typical test run involved a fish sample placed on the balance pan, and impacted onto its dorsal area by a small plastic hammer released from a specific height.…”

Section: Methodsmentioning

confidence: 99%

“…Thus, unless a simplified system model is deemed sufficient (as in Grigorakis and Dimogianopoulos ), it is difficult to combine the complex or partially known physics‐based models of each component into one accurate system model. Therefore, the approach in Dimogianopoulos and Grigorakis () based on the use of system response data for identifying system models of a particular form was also used in this study. The simplest such form is that of the output‐only linear stochastic autoregressive (AR) representations:

\begin{matrix} y [t] = \sum_{i = 1}^{n_{y}} θ_{i} \cdot p_{i} [t] + e [t] \\ e [t] - N I D (0, σ_{e}^{2}) \end{matrix}

with t designating the normalized discrete time, and y [ t ], e [ t ] the output and one step ahead prediction error (or residual, assumed to be a zero mean uncorrelated sequence with variance

σ_{e}^{2}

) signals, respectively.…”

Section: Methodsmentioning

confidence: 99%

“…). The low sensitivity during the early shelf life is the weak point for many of these applications, as commented in Dimogianopoulos and Grigorakis ().…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Nondestructive Textural Assessment of Fish Freshness: A Stochastic Model‐Based Approach Robust to Fish Size Variations

Dimogianopoulos

Grigorakis

2014

Journal of Texture Studies

View full text Add to dashboard Cite

This paper aims at proposing a nondestructive scheme for assessing freshness of whole raw fish of various sizes/weights, using a stochastic model-based fault diagnosis framework. Unlike most alternatives, the scheme detects early postmortem texture alterations linked to freshness reductions, thus facilitating the accurate estimation of high-quality shelf life of fish. The method involves the experimental testing of fish samples via vibration-like procedures, and the evaluation of differences between a sample of unknown freshness and a whole group of fresh fish of possibly different sizes/weights. Freshness degree is reliably concluded via statistical decision-making hypothesis tests, which also quantify the uncertainty of the process. Eleven whole raw sea bass (Dicentrarchus labrax) of various sizes/weights and freshness degrees were tested, with freshness results correctly assessed as confirmed by organoleptic freshness evaluation. PRACTICAL APPLICATIONSThe proposed diagnostic scheme introduces a practical, nondestructive method for assessing fish freshness by evaluating texture alterations based on a stochastic model-based fault diagnosis framework. The scheme in its current stage of development detects differences between a sample of unknown freshness and a group of fresh fish, without requiring similar sizes/weights of the tested fish samples. Freshness degree is reliably concluded via statistical decision-making hypothesis tests, which also quantify the uncertainty of the process. These characteristics make it ideal for assessing freshness in fisheries and aquaculture industry, with tested fish samples of unknown freshness being compared with a (potentially expanding) database of "benchmark" fresh fish.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

\begin{matrix} y [t] = \sum_{i = 1}^{n_{y}} θ_{i} \cdot p_{i} [t] + e [t] \\ e [t] - N I D (0, σ_{e}^{2}) \end{matrix}

with t designating the normalized discrete time, and y [ t ], e [ t ] the output and one step ahead prediction error (or residual, assumed to be a zero mean uncorrelated sequence with variance

σ_{e}^{2}

) signals, respectively.…”

Section: Methodsmentioning

confidence: 99%

“…). The low sensitivity during the early shelf life is the weak point for many of these applications, as commented in Dimogianopoulos and Grigorakis ().…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations