Keeping It Fresh: New AI-based Strategy Can Assess the Freshness of Beef Samples

Victoria D. Doty

Researchers incorporate spectroscopy and deep discovering in an effective technique for detecting spoiled meat.

Researchers at Gwangju Institute of Science and Technologies, Korea, incorporate an affordable spectroscopy technique with artificial intelligence to create a new way of evaluating the freshness of beef samples. Their approach is remarkably quicker and extra price tag-helpful than regular methods when maintaining a fairly superior accuracy, paving the way for mass-developed units to detect spoiled meat each in the industry and at house.

Picture credit history: Pixabay (Cost-free Pixabay license)

Whilst beef is 1 of the most consumed food items all-around the globe, consuming it when it’s past its primary is not only unsavory, but also poses some serious wellbeing pitfalls. Unfortunately, accessible techniques to test for beef freshness have numerous shortcomings that hold them from getting valuable to the community. For example, chemical analysis or microbial populace evaluations consider much too much time and need the skills of a skilled. On the other hand, non-harmful methods dependent on around-infrared spectroscopy need expensive and complex equipment. Could artificial intelligence be the important to a extra price tag-helpful way to evaluate the freshness of beef?

At Gwangju Institute of Science and Technologies (GIST), Korea, a crew of experts led by Associate Processors Kyoobin Lee and Jae Gwan Kim have formulated a new method that combines deep discovering with diffuse reflectance spectroscopy (DRS), a fairly affordable optical technique. “Unlike other kinds of spectroscopy, DRS does not need complicated calibration instead, it can be made use of to quantify part of the molecular composition of a sample working with just an affordable and easily configurable spectrometer,” clarifies Lee. The findings of their examine are now posted in Foodstuff Chemistry.

To determine the freshness of beef samples, they relied on DRS measurements to estimate the proportions of distinctive types of myoglobin in the meat. Myoglobin and its derivatives are the proteins mainly accountable for the coloration of meat and its alterations in the course of the decomposition process. Having said that, manually changing DRS measurements into myoglobin concentrations to at last determine on the freshness of a sample is not a very accurate strategy—and this is wherever deep discovering will come into enjoy.

Convolutional neural networks (CNN) are greatly made use of artificial intelligence algorithms that can study from a pre-categorised dataset, referred to as ‘training set,’ and discover hidden designs in the knowledge to classify new inputs. To practice the CNN, the scientists gathered knowledge on seventy eight beef samples in the course of their spoilage process by consistently measuring their pH (acidity) along with their DRS profiles. After manually classifying the DRS knowledge dependent on the pH values as ‘fresh,’ ‘normal,’ or ‘spoiled,’ they fed the algorithm the labelled DRS dataset and also fused this info with myoglobin estimations. “By giving each myoglobin and spectral info, our properly trained deep discovering algorithm could accurately classify the freshness of beef samples in a make any difference of seconds in about ninety two{394cb916d3e8c50723a7ff83328825b5c7d74cb046532de54bc18278d633572f} of scenarios,” highlights Kim.

Apart from its accuracy, the strengths of this novel method lie in its velocity, very low price tag, and non-harmful character. The crew thinks it may well be feasible to create little, moveable spectroscopic units so that absolutely everyone can easily evaluate the freshness of their beef, even at house. Also, equivalent spectroscopy and CNN-dependent procedures could also be extended to other products, these types of as fish or pork. In the foreseeable future, with any luck, it will be less complicated and extra accessible to detect and avoid questionable meat.


Authors: Sungho Shin (one), Youngjoo Lee (2), Sungchul Kim (2), Seungjun Choi (one), Jae Gwan Kim (2) Kyoobin Lee (one)

Title of initial paper:       Quick and non-harmful spectroscopic approach for classifying beef freshness working with a deep spectral network fused with myoglobin info

Journal: Foodstuff Chemistry

DOI: ten.1016/j.foodchem.2021.129329


  • Faculty of Built-in Technologies, Gwangju Institute of Science and Technologies (GIST)
  • Department of Biomedical Science & Engineering, Gwangju Institute of Science and Technologies (GIST)

About Gwangju Institute of Science and Technologies (GIST)

Gwangju Institute of Science and Technologies (GIST) is a investigate-oriented college located in Gwangju, South Korea. One particular of the most prestigious colleges in South Korea, it was founded in 1993. The college aims to build a powerful investigate environment to spur advancements in science and technological innovation and to advertise collaboration involving international and domestic investigate courses. With its motto, “A Very pleased Creator of Long run Science and Technologies,” the college has consistently obtained 1 of the greatest college rankings in Korea.

Internet site:

About the authors

Kyoobin Lee is an Associate Professor and Director of the AI laboratory at GIST. His group is developing AI-dependent robot vision and deep discovering-dependent bio-professional medical analysis techniques. Ahead of becoming a member of GIST, he acquired a PhD in Mechatronics from KAIST and completed a postdoctoral teaching method at Korea Institute of Science and Technologies (KIST).

Jae Gwan Kim is an Associate Professor at the Department of Biomedical Science and Engineering at GIST due to the fact 2011. His present investigate subject areas include brain stimulation by transcranial ultrasound, anesthesia depth checking, and screening the phase of Alzheimer’s sickness via brain functional connectivity measurements. Ahead of becoming a member of GIST, he completed a postdoctoral teaching method at the Beckman Laser Institute and Health-related Clinic at UC Irvine, United states of america. In 2005, he obtained a PhD in Biomedical Engineering from a joint method involving the College of Texas at Arlington and the College of Texas Southwestern Health-related Heart at Dallas, United states of america.

Next Post

AlphaEvolve: A Learning Framework to Discover Novel Alphas in Quantitative Investment

Alphas are inventory prediction designs making triggers to obtain or sell shares. In this area, current AI strategies surpass human-created alphas. Present strategies benefit from only brief-phrase capabilities or are extremely sophisticated. A new study paper suggests a novel course of alphas that blend the strengths of current kinds. They […]

Subscribe US Now