INTRODUCTION:

Spices have had a significant impact on human history, shaping culinary customs, trading routes, and cultural interactions for ages. In addition to their capacity to improve the taste and scent of food, spices have been highly esteemed for their therapeutic attributes and ability to preserve food. The use of spices goes beyond simple enjoyment of taste, including a diverse range of cultural, economic, and health importance. Nevertheless, the spice sector has recently been confronted with a substantial obstacle - the adulteration of spices^1,2.

Significance of spices:

Spices serve as the fundamental element of culinary arts, imparting richness, intricacy, and distinctiveness to foods.

Spices, such as cardamom, chili peppers, cinnamon, and cumin, provide distinct tastes that elevate ordinary dishes into exceptional culinary experiences. Spices have played a crucial role in shaping cultural identity and legacy. These agents have a pivotal position in customary gastronomies, joyous festivities, and ceremonial practices worldwide. The use of certain spices often mirrors the historical background, geographical characteristics, and trade routes of a given area, resulting in a varied and interrelated gastronomic realm.Numerous spices possess therapeutic characteristics that have been used for millennia in diverse traditional treatment methods^3,4. Turmeric, for example, includes curcumin, which is recognized for its anti-inflammatory and antioxidant characteristics. Spices have historically been used for the purpose of alleviating illnesses, enhancing the body's defence mechanisms, and fostering general health and prosperity^5,6. Traditionally, spices were used not only for their taste but also for their capacity to safeguard food^7,8. Their antibacterial capabilities provided protection against spoilage and foodborne infections, particularly in periods when refrigeration was unavailable. The spice trade has had a significant influence on worldwide commercial endeavors for countless years. The availability to and control over spice routes have had a profound impact on the birth and collapse of whole civilizations and empires. Currently, the spice sector continues to play a substantial role in the economy of several nations, supporting the lives of millions of people^9,10. Spices has the capacity to be transformed into substances that may serve as preservatives or therapeutic treatments for oral ailments^11,12.

The issue of adulteration:

Although spices are very valuable, the spice business has a significant obstacle i.e., the adulteration. Adulteration refers to the act of adding substandard or toxic ingredients to spice items, which undermines their quality and safety^13–15. This practice presents significant risks to the well-being of the general population, the stability of the economy, and the preservation of cultural and gastronomic customs. Adulterated spices may include toxic compounds such as chemicals, synthetic dyes, and impurities that provide health hazards. Ingesting adulterated spices may result in allergies, gastrointestinal complications, or even chronic health conditions^16,17. Adulteration compromises the economic integrity of the spice sector. Genuine and superior spices demand higher cost, yet, unethical activities mislead customers, damage reputations, and impede equitable commerce.Cultural dilution occurs when the authenticity of spices is compromised, resulting in a decrease in their cultural value. The authenticity of traditional recipes, which have been handed down through generations, is compromised when prepared using contaminated components.The widespread occurrence of spice adulteration undermines customer faith in the sector. This not only affects companies but also erodes the faith customers have in the safety and genuineness of the spices they buy¹⁸.

Cardamom: The case study:

Cardamom (Elettaria cardamom), often known as the "Queen of Spices," serves as a majestic symbol of the significant role spices play in the worlds of cuisine and culture. Cardamom, native to the fertile terrains of the Indian subcontinent, has become an essential component in international cuisines, celebrated for its alluring fragrance, comforting, tangy taste, and multitude of health advantages. The flexibility of this ingredient is shown in a wide range of recipes, including both sweet and savory ones, such as fragrant chai teas and deep, intricate curries. In addition to its culinary uses, cardamom is highly valued for its medical characteristics, which include promoting digestion, improving respiratory health, and acting as a natural mood booster¹⁹.

Nevertheless, the magnificence of cardamom is endangered by the widespread problem of adulteration, a behaviour that undermines its integrity and genuineness. A popular kind of adulteration includes replacing genuine cardamom seeds with those of Citrus sinensis, which is usually referred to as sweet orange. The desiccated seeds of Citrussinensis, while superficially resembling cardamom, do not possess the peculiar and intricate taste character that makes cardamom very desirable²⁰. This deceitful method not only weakens the authenticity of cardamom but also presents possible health hazards, as the contaminated spice may introduce unknown components and allergies into culinary preparations. Various adulterants related to cardamom are shown in Table 1.

Table 1: List of adulterants of cardamom spice.

Spice

In raw cardamom

In powdered cardamom

Biological adulterants

Physical adulterants

Adulteration in powdered form

Cardamom

(Elettaria cardamomum)

Seeds of Citrus sinensis (sweet orange).

Seed mixed with Ammonium aromaticum or A. subulatum

Small pebbles and unroasted coffee seeds

Cardamom powder can be adulterated by material powdered to similar size

To maintain the grandeur of cardamom and ensure the safety of customers, collaborative efforts are necessary across the spice sector. To effectively address the issue of cardamom adulteration, it is crucial to implement rigorous quality control methods, ensure transparency in supply chains, and establish strong regulatory frameworks. In addition, providing customers with information about the distinctive qualities of real cardamom and promoting their ability to make well-informed decisions will help foster a stronger and more genuine spice industry. As we explore the complex world of global cuisine, it is important to acknowledge and tackle the problems caused by food adulteration in order to preserve the esteemed status of spices in our culinary traditions.

Lack of digital method for authentication of spices:

The authentication of spices, which is crucial for guaranteeing their quality and safety, suffers from a significant lack of strong digital solutions. Conventional authentication techniques, while respected for their longevity, can require a lot of effort, are based on personal skill, and are susceptible to mistakes made by humans. The sensory assessment, predicated on the acumen of humans to evaluate attributes such as fragrance, taste, and appearance, is profoundly subjective and might diverge across disparate assessors. Furthermore, relying just on visual examination may not be sufficient to identify intricate types of adulteration, particularly in the case of powdered or processed spices. Chemical analysis, a traditional approach, entails intricate laboratory protocols that are both time-consuming and costly. Moreover, conventional approaches may lack the necessary sensitivity to detect more recent types of adulteration.

The absence of a comprehensive and effective digital authentication technique intensifies these issues. A reliable digital method, including techniques such as method based on the artificial intelligence (AI) and machine learning (ML) might provide quick and unbiased results, providing a degree of accuracy and dependability that surpasses existing procedures. Digital authentication not only expedites the procedure but also improves precision, allowing for the identification of slight tampering or pollution that may be overlooked with traditional methods. Furthermore, digital techniques have the benefit of establishing a clear and verifiable system for tracking the supply chain, which enhances customer trust in the genuineness and source of the spices they buy.

Previous research related to ML based identification of herbal entities:

In 2023, authors conducted intriguing research on the identification of cardamom using a ML model and compared three transfer learning models to determine the most effective approach²¹. Additionally, the wehas also successfully identified raw C. annum using an ML model, achieving acceptable accuracy²². In 2021, Liu et al. developed an automatic AutoML model for identifying 315 varieties of traditional chinese medicine with impressive accuracy²³. Lv et al., (2021) developed the ConvNeXt ML model for identifying chinese herbal medicine²⁴. Similarly, Rao et al. (2023) developed a Convolutional Neural Networks (CNN) model for identifying Ayurvedic plants based on leaf images²⁵. Furthermore, Kumar and his team developed an automatic recognition model for identifying therapeutic plants, achieving an accuracy of 98.98%²⁶. Dadioset al., (2019) also contributed to the field by developing an ML model with accuracies of 94.5% and 93.3% for identifying Philippine plants²⁷. These methods have shown significant potential, particularly in identifying herbs through leaves or other plant parts.While research in the area of identification of spices and raw spices remains limited, the potential for ML models in identifying spices or raw herbs is promising.

The results of this study show great potential in the current pharmaceutical sector, which places significant importance on the need for superior and uncontaminated constituents. Enhancing the authenticity of herbal medications and spices may be achieved by integrating ML, AI, and conventional techniques^28–30. This not only ensures the genuineness of pharmaceutical goods that include herbs and spices, as well as their extracts, but also minimizes the possible health risks linked to contaminated herbs. Moreover, the precise quantification of bioactive compounds like capsaicin, facilitated by advanced methods like high-performance liquid chromatography (HPLC), is crucial in the development of pharmaceutical solutions that are uniform and reliably efficacious^31–33. Hence, this interdisciplinary approach has the capacity to enhance the pharmaceutical industry's efforts in ensuring the safety, efficacy, and authenticity of therapeutic therapies derived from herb/spice-based constituents.

MATERIAL AND METHODS:

New approach comprising of machine learning model for the authentication of spices:

AI and ML are powerful technologies that may greatly impact several sectors, including spice authentication. AI and ML, specifically transfer learning models, have been more important in guaranteeing the genuineness and excellence of spices in recent times^34–37. Within the spice sector, the widespread problem of adulteration presents considerable obstacles. Therefore, the use of sophisticated technology has become essential. Transfer learning (TL), a subfield of ML, enables the use of models that have been previously trained on one job to be applied to another activity, providing a potent tool for verifying the authenticity of spices. This integration of state-of-the-art technology not only overcomes the shortcomings of conventional authentication techniques but also allows for quick, precise, and adaptable solutions in identifying adulteration, guaranteeing the authenticity of spice supply chains, and inspiring trust in customers. Exploring the complex realm of AI and ML in spice authentication reveals the capacity of these technologies to enhance the benchmarks of quality control, traceability, and transparency in the spice sector.

The process of creating a model to authenticate spices using AI and ML, specifically using TL, entails a sequence of well-defined procedures^38,39. Presented below is an elaborated step wise procedure delineating the essential steps involved in the construction of such a model:

· Identification of unique feature: Every spice and its adulterants have some unique feature(s) morphologically and microscopic/histological examination. These unique features act as inputs for development and training of ML models.

· Generation of dataset: Compile an extensive collection of spice samples, including both genuine and adulterants sample. The dataset should be inclusive, including a wide range of spice varieties, sources, and possible adulterants and assigning labels to each sample in the collection to indicate its legitimacy⁴⁰.

· Data pre-processing: Datasets are to be processed by handling missing numbers, outliers, and irregularities in the dataset. Standardization and normalization techniques will be applied to ensure that characteristics are brought to a uniform scale. If needed, expand the dataset to enhance its size and variability, particularly when dealing with a restricted number of samples through the process of data augmentation⁴¹.

· Choosing the most appropriate model:Selecting an appropriate ML framework depending on the intricacy of the challenge is one of the most important steps. CNNs are often used for image-based applications such as spice authentication because to their high effectiveness. Utilizing of a pre-existing model via TL to capitalize on the information acquired from a similar assignment will be can become important part^42–44.

· Transfer learning: Choosing a pre-existing model (e.g., ResNet, VGG, Inception) that has undergone training on a substantial dataset for a distinct but interconnected goal, such as identifying imagesis an important step. Structure of the model will be adapted to accommodate the spice authentication job by either altering the existing final layers or including new ones. Pre-trained model are utilized as a feature extractor by eliminating its final layers. Significant characteristics i.e., unique features from the collection by running the photos through the altered model are applied for further optimization and selection of suitable model^45–47.

· Training the model:To enhance the model performance for spice authentication, the entire dataset needs to be split into separate training and validation sets. The weights will then be fine-tuned on the training set, and regularization techniques will be applied to prevent overfitting.

· Hyperparameter adjustment: Model performance needs to be optimized by fine-tuning hyperparameters, including the learning rate, batch size, and optimization techniques.

· Evaluation of the model: The model performance will be evaluated through the validation set by measuring measures like as accuracy, precision, recall, and F1-score^48,49.

· Testing and deployment: Model's generalization performanceshould be accessed by validating it on a separate test set.After verifying the model's correctness and reliability, proceed to deploy it for real implementation in spice authentication.

By adhering to these procedures, the creation of an AI and ML model for spice authentication may result in a sturdy and dependable solution, therefore bolstering quality control in the spice sector.

Figure 1: ML model for authentication of spices.

RESULTS AND DISCUSSION:

Development of ML based digital model for the identification of cardamom:

As a representative case study, cardamom, a widely used spice, often encounters the problem of adulteration. This case study examines the use of machine learning to accurately detect cardamom and its possible adulterants. This is achieved by constructing datasets and creating a strong identification model.

Identification of unique features:

Identification of unique morphological features is essential for development of comprehensive digital strategy to recognizing cardamom and its possible biological adulterants, by using a novel ML model. Significantly, the seeds of Citrus sinensis and Amomum subulatum are the notable biological adulterants of cardamom. By using morphological and organoleptic traits, it is possible to distinguish genuine cardamom from its adulterants.

Authentic cardamom may be visually identified by its unique color range, which spans from vivid green to a delicate yellow shade. On the other hand, Amomum subulatum, which is another kind of biological adulterant, exhibits a wide range of hues, particularly varying from deep red to brown (as shown in Fig. 2).

The seeds of Citrus sinesis may be further distinguished based on their surface features. Genuine cardamom seeds have distinct ridges on their surfaces, which contribute to a textured look. This tactile characteristic functions as a means of identifying via touch, strengthening the visual indicators for the purpose of authentication. Conversely, the seeds of Citrus sinensis have a much smoother surface, without the distinct ridges seen in authentic cardamom (shown in Fig.2). These visual differences act as a crucial signal in determining the genuineness of cardamom, enabling quick and instinctive detection²⁰.The difference in surface texture improves the dependability of the optical and tactile identification techniques. Ultimately, integrating the knowledge gained from the recently created ML model with conventional visual and tactile techniques establishes a strong authentication framework for cardamom and its biological adulterants. An in-depth analysis of color variations and surface features guarantees a thorough method for protecting the genuineness of this highly valued spice, so strengthening customer trust and upholding the integrity of the spice sector^19,20.

Dataset creation:

A total of 20000 images were gathered in different form of PNG, JPG etc., to develop the cardamom datasets. Glimpse of datasets of the cardamom datasets is shown in Fig. 3.

Figure 3: Representative datasets of the cardamom.

Pre-processing:

The images of datasets underwent pre-processing, were then converted to the JPEG format, and were reduced to dimensions of 224 by 224 pixels. The dataset was divided into training and testing sets at a ratio of 70:30. In order to enhance accuracy, data augmentation approaches were used. This included incorporating a variety of instances during training by using the Keras "Image Data Generator" API.

Development of the CNN-TL MobileNet model:

This study focuses on the process of fine-tuning previously developed model for the identification of cardamom i.e., Mobile Net. Developed models were first trained using the extensive ImageNet dataset. By implementing the training on other model, we translate the knowledge and expertise gained from ImageNet to our specific cardamom dataset. The final constituted levels of models, include Softmax, fully connected, and classification layers, with novel fully connected layers that correspond to the number of classes in the fresh training dataset. The modified and reconstituted Mobile Netmodel was then trained using the cardamom pictures. The models were fine-tuned and compared based on different parameters and accuracy.

Model performance analysis:

The evaluation of the developed model was analyzed bymeasuring accuracy, Re-call, F1-score, time elapsed in model development and results are shown in Table 2.

Table 2: Model performance results.

Model	Average Recall	Average F1-score	Time Elapsed	Model Size (MB)
M2	0.9230	0.9318	751	98

After training the Mobile Net model on the image datasets of the cardamom, the developed pre-trained model Mobile Net, it attained the satisfactory level of accuracy, reaching 95.5% (Fig 4) in 10 epochs. Despite its slightly lower accuracy in comparison with the identification of cardamom, the Mobile Net model was further employed for its computational efficiency, making it suitable for low-power devices. Our analysis showed that transfer learning is crucial for models with limited datasets.

Fig. 4: Training accuracy for the developed CNN MobileNet model for the identification of the cardamom.

The developed Mobile Net model is effective in identifying cardamom and its adulterants, especially on devices with low computational capabilities. However, further improvements are needed for handling images with complex backgrounds. The confusion matrix of the Mobile Net model displayed a prominent pattern, with the greatest number of correct positive predictions occurring notably in the recognition of cardamom and non-cardamom (adulterants) datasets (Fig. 5). Consequently, the majority of positive predictions in the model's classification findings were situations when cardamom was successfully identified.

Fig. 5: Confusion matrix of the CNN MobileNet model for the identification of cardamom and its adulterants.

The Mobile Net model exhibited remarkable competence in identifying cardamom, highlighting its usefulness in distinguishing this specific spice. The research showed that the CNN-MobileNet model demonstrated the essential attributes for effective and precise detection of cardamom and its adulterants. The model exhibited exceptional accuracy, recall, and F1-score (Table 2), rendering it suitable for expedited identification tasks with constrained processing resources and storage²¹.

As previously stated, there has been relatively little studies on identifying raw spices and herbs using an ML model. As the previous research of our group on a CNN-based ML model for identifying cardamom and chilli, as well as Chen et al., (2023) Auto-ML model for 315 traditional Chinese medications, represent significant advances in the field of herb and spice identification using ML techniques^21–23. In this study, the CNN MobileNet model achieved an impressive accuracy in identification of the cardamom and its adulterants which pave the way for the development of more such model for the spices which are vulnerable with the problem of the adulterations.

Industrial Techniques for Identification of Adulterants in Raw and Powdered Cardamom:

Identifying adulterants in raw and powdered forms of cardamom in the industrial setting require a comprehensive methodology that integrates cutting-edge technology with conventional approaches to guarantee precision and dependability^50,51.

Hyper-spectral imaging methods may be used to get comprehensive spectral information from the surface of raw cardamom. This enables the identification of genuine cardamom and distinguishes it from possible contaminants such as seeds of Citrus sinensis and Amomum subulatum. These strategies use the distinct spectrum characteristics of each chemical to provide a complete and discriminative dataset for analysis⁵².

When working with powdered cardamom, the difficulties become more complex. The fourier transform infrared (FTIR) spectroscopy is very helpful due to its capability to evaluate the molecular makeup of substances. FTIR spectroscopy enables the fast and non-destructive identification of cardamom authenticity by analyzing its infrared spectra and comparing them to those of typical adulterants such starch or other plant-derived fillers. HPLC may be used to isolate and measure various constituents in powdered samples, verifying that the ratios correspond to the anticipated composition of genuine cardamom. Simultaneously, microscopic examination may uncover textural discrepancies in the powder, facilitating the detection of extraneous particles^53,54.

Furthermore, integration of these sophisticated technologies with conventional approaches improves the reliability of the identification process. Conducting visual examinations to identify differences in color and texture, together with analysing the flavor, offers further levels of verification. To ensure the purity and quality of cardamom in the spice sector, it is necessary to use a wide range of advanced procedures that take use of technical breakthroughs, while still recognizing the importance of traditional observations. The combination of these methodologies not only guarantees the precise identification of adulterants but also offers a basis for preserving the authenticity of the spice supply chain^55–57.

SUMMARY AND FUTURE SCOPE:

To summarize, this investigation into the domain of spices, with specific emphasis on cardamom, uncovers the significant influence spices have had on human history, culture, and trade. In addition to their culinary contributions, spices possess cultural, economic, and health value. Not only can they enhance the flavour and perfume of food, but they also possess healing properties and historical significance. Adulteration represents a significant issue to the spice industry, since it damages the authenticity of spices, presents health hazards, and weakens economic and cultural factors. Cardamom, referred to as the "Queen of Spices," embodies the diverse and significant role that spices play in worldwide cuisine and cultural traditions. The spice's desirability stems from its ability to be used in a wide range of sweet and savory foods, as well as its therapeutic attributes. Regrettably, the extensive adulteration of cardamom, including the replacement of legitimate seeds with those from C. sinensis, poses a danger to its genuineness and the welfare of customers.

In order to tackle these difficulties, a comprehensive strategy has been proposed. The study provides an overview of the creation of a machine learning model, especially a CNN-MobileNet model based on transfer learning, with the purpose of authenticating cardamom. Trained on a dataset of 20,000 cardamom photos, this model has good accuracy, recall, and F1-score, indicating its ability in discriminating real cardamom from possible adulterants.

Furthermore, there has been a focus on using conventional procedures that rely on sight and touch to authenticate cardamom and identify any biological adulterants present. Visual clues, such as changes in color and surface characteristics, together with tactile distinctions in seed texture, provide concrete means to distinguish genuine cardamom from substitutes like seeds of C. sinensis and A. subulatum.

Advanced methods, including hyperspectral imaging, FTIR spectroscopy, HPLC, and microscopic analysis, have been suggested for detecting adulterants in raw and powdered cardamom in industrial settings. These approaches provide a resilient and all-encompassing strategy to guaranteeing the integrity and excellence of cardamom goods.

In the future, the use of digital techniques, including AI and ML, shows significant potential for improving spice authentication procedures. Additional progress in ML models, perhaps integrating more intricate structures and algorithms, might enhance accuracy and efficiency to a greater extent. Furthermore, the incorporation of blockchain technology for clear supply chain management and the creation of portable, user-friendly devices for immediate verification provide promising opportunities for future investigation.

To summarize, the complex network of difficulties related to spice authentication necessitates a comprehensive strategy that integrates conventional techniques, cutting-edge technology, digital advancements, and development of digital platform like mobile application. The continuous endeavour to achieve authenticity in the spice industry is crucial not only for maintaining economic sustainability but also for safeguarding cultural heritages and guaranteeing the welfare of customers. The integration of conventional knowledge with state-of-the-art technology is set to influence the trajectory of spice authentication, cultivating a spice sector that is open, reliable, and genuine to its origins.

FUNDING:

This study was funded by Department of Science and Technology, Government of India under Science and Engineering Research Board (SERB), Core Research Grant (CRG) Scheme (Project No. CRG/2018/00425) to Prof. Ashish Baldi.

CONFLICT OF INTEREST:

The authors have no conflicts of interest regarding this research.

ACKNOWLEDGMENTS:

The authors would like to thank Department of Science and Technology for providing the funds for carrying out this research work. Authors also thanks Department of Pharmaceutical Science and Technology, MRSPTU Bathinda for providing the infrastructure for carrying out the study.

REFERENCES:

1. Uhl SR. Handbook of Spices, Seasonings and Flavorings. CRC Press. 2000. https://doi.org/10.1201/9781003040569

2. Sharangi AB, Acharya SK. Spices in India and beyond: the origin, history, tradition and culture. In: Indian Spices, Edited by Sharangi AB. Springer, 2018; 1–11. https://doi.org/10.1007/978-3-319-75016-3_1

3. Peter KV, Babu KN. Introduction to herbs and spices: medicinal uses and sustainable production. In: Handbook of Herbs and Spices, Edited by Peter KV. Elsevier, 2012; 2nd ed: 1–16. https://doi.org/10.1533/9780857095688.1

4. Sharangi AB, Pandit MK. Supply chain and marketing of spices. In: Indian Spices, Edited by Sharangi AB. Springer, 2018; 341–57. https://doi.org/10.1007/978-3-319-75016-3_12

5. Singh PA, Bajwa N, Baldi A. A comparative review on the standard quality parameters of turmeric. Indian Journal of Natural Products. 2021; 35(1): 2-8. http://dx.doi.org/10.5530/ijnp.2021.1.2

6. Niranjan A, Prakash D. Chemical constituents and biological activities of turmeric (Curcuma longa l.)-A review. Journal of Food Scienceand Technology. 2008; 45(2): 109-16.

7. Duke JA. CRC Handbook of Medicinal Spices. CRC press, 2002.https://doi.org/10.1201/9781420040487

8. Ravindran PN. The Encyclopedia of Herbs and Spices. CABI, 2017.

9. Gupta M. Pharmacological properties and traditional therapeutic uses of important Indian spices: A review. International Journal of Food Properties. 2010; 13(5): 1092–1116. https://doi.org/10.1080/10942910902963271

10. Singh NA, Kumar P, Kumar N. Spices and herbs: Potential antiviral preventives and immunity boosters during COVID‐19. Phytotherapy Research. 2021; 35(5): 2745–57. https://doi.org/10.1002/ptr.7019

11. Shankar S, Gopinath P, Roja E. Role of spices and herbs in controlling dental problems. Research Journal of Pharmacology and Pharmacodynamics. 2022; 14(1): 23–28. http://dx.doi.org/10.52711/2321-5836.2022.00004

12. Parvathi P, Geetha R V. Spices and oral health. Research Journalof Pharmacy and Technology. 2014; 7(2): 235–237.

13. Sasikumar B, Swetha VP, Parvathy VA, Sheeja TE. Advances in adulteration and authenticity testing of herbs and spices. In: Advances in Food Authenticity Testing, Edited by Gerard Downey. Elsevier, 2016; 585–624. https://doi.org/10.1016/B978-0-08-100220-9.00022-9

14. Panda H. Handbook on Spices and Condiments (Cultivation, Processing and Extraction). Asia Pacific Business Press Inc., 2010.

15. Mohiuddin AK. Health hazards with adulterated spices: Save the “onion tears”. Asian Journal of Research in Pharmaceutical Science. 2020; 10(1): 21–25. http://dx.doi.org/10.5958/2231-5659.2020.00005.3

16. Beniwal A, Khetarpaul N. Knowledge of consumers regarding the nature and extent of adulteration of Indian foods. Nutrition and Health. 1999; 13(3): 153–60. https://doi.org/10.1177/026010609901300303

17. Velázquez R, Rodríguez A, Hernández A, Casquete R, Benito MJ, Martín A. Spice and herb frauds: Types, incidence, and detection: The state of the art. Foods. 2023; 12(18): 1-38.

18. Osman AG, Raman V, Haider S, Ali Z, Chittiboyina AG, Khan IA. Overview of analytical tools for the identification of adulterants in commonly traded herbs and spices. Journal of AOAC International. 2019; 102(2): 376–385.

19. Battaglia S. Cardamom [Homepage on the Internet]. 2019 [cited on 2024; Mar 15; 2024]; Available from: http://www.perfectpotion.com.au/.

20. Govindarajan VS, Narasimhan S, Raghuveer KG, Lewis YS, Stahl WH. Cardamom—Production, technology, chemistry, and quality. Critical Reviews in Food Science & Nutrition. 1982; 16(3): 229–326. https://doi.org/10.1080/10408398209527337

21. Naman S, Sharma S, Kumar M, Kumar M, Baldi A. Developing a CNN-based machine learning model for cardamom identification: A transfer learning approach. Latin American Journal of Pharmacy. 2023; 42(6): 565–74.

22. Naman S, Sharma S, Baldi A. Machine learning based identification of spices: A case study of chilli. Latin American Journal of Pharmacy: A Life Science Journal. 2023; 42(10): 248–61.

23. Chen W, Tong J, He R, Lin Y, Chen P, Chen Z, Liu X. An easy method for identifying 315 categories of commonly-used Chinese herbal medicines based on automated image recognition using AutoML platforms. Informatics in Medicine Unlocked. 2021; 25: 100607. https://doi.org/10.1016/j.imu.2021.100607

24. Miao J, Huang Y, Wang Z, Wu Z, Lv J. Image recognition of traditional Chinese medicine based on deep learning. Frontiers in Bioengineering and Biotechnology. 2023; 11: 1-10. https://doi.org/10.3389/fbioe.2023.1199803

25. Rao MS, Kumar SP, Rao KS. A Methodology for Identification of ayurvedic plant based on machine learning algorithms. International Journal of Computing and Digital Systems 2023; 14(1): 10233–41. http://dx.doi.org/10.12785/ijcds/140196

26. Shailendra R, Jayapalan A, Velayutham S, Baladhandapani A, Srivastava A, Kumar Gupta S, Kumar M. An IoT and machine learning based intelligent system for the classification of therapeutic plants. Neural Process Letter. 2022; 54(5): 4465–93. https://doi.org/10.1007/s11063-022-10818-5

27. Luna RG De, Rosales MA, Dadios EP. Classification of philippine herbal plants via leaf using different machine learning algorithms. Journal of Computational Innovations and Engineering Applications. 2019; 4(1): 29–34.

28. Tahilani P, Swami H, Goyanar G, Tiwari S. The era of artificial intelligence in pharmaceutical industries-A review. Indian Journal of Pharmacy & Drug Studies. 2022; 1(2): 47–50.

29. Pillai S, Chakraborty J. A study to assess the knowledge regarding food adulteration among home makers regarding food safety standards in selected rural community. Asian Journal of Nursing Education and Research. 2017; 7(1): 77–78. http://dx.doi.org/10.5958/2349-2996.2017.00016.7

30. Bendre S, Shinde K, Kale N, Gilda S. Artificial intelligence in food industry: A current panorama. Asian Journal of Pharmacy and Technology. 2022; 12(3): 242–250. http://dx.doi.org/10.52711/2231-5713.2022.00040

31. Patel SS, Shah SA. Artificial intelligence: Comprehensive overview and its pharma application. Asian Journal of Pharmacy and Technology. 2022; 12(4): 337–348. http://dx.doi.org/10.52711/2231-5713.2022.00054

32. Patel AI, Khunti PK, Vyas AJ, Patel AB. Explicating artificial intelligence: Applications in medicine and pharmacy. Asian Journal of Pharmacy and Technology. 2022; 12(4): 401–406. http://dx.doi.org/10.52711/2231-5713.2022.00061

33. Kulkarni RR, Pawar PS. Artificial intelligence in pharmacy. Asian Journal of Pharmacy and Technology. 2023; 13(4): 304–306.http://dx.doi.org/10.52711/2231-5713.2023.00054

34. Došilović FK, Brcic M, Hlupić N. Explainable artificial intelligence: A survey. In: 2018 41stInternational Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO). IEEE, 2018; 210–5. https://doi.org/10.23919/MIPRO.2018.8400040

35. Jiang F, Jiang Y, Zhi H, Dong Y, Li H, Ma S, Wang Y, Dong Q, Shen H, Wang Y. Artificial intelligence in healthcare: Past, present and future. Stroke and Vascular Neurology. 2017; 2(4): 230-43. https://doi.org/10.1136/svn-2017-000101

36. Che Soh A, Yusof UK, Radzi NFM, Ishak AJ, Hassan MK. Classification of aromatic herbs using artificial intelligent technique. Pertanika Journal of Science & Technology. 2017; 25. 119-25. http://www.pertanika.upm.edu.my/

37. Prieto A, Atencia M, Sandoval F. Advances in artificial neural networks and machine learning. Neurocomputing. 2013; 121: 1–4. https://doi.org/10.1016/j.neucom.2013.01.008

38. Sharma S, Naman S, Dwivedi J, Baldi A. Artificial intelligence-based smart identification system using herbal images: Decision making using various machine learning models. In: Applications of Optimization and Machine Learning in Image Processing and IoT. Chapman and Hall/CRC, 123–55. https://doi.org/10.1201/9781003364856

39. Singh PA, Bajwa N, Naman S, Baldi A. A review on robust computational approaches based identification and authentication of herbal raw drugs. Letter Drug Design and Discovery. 2020; 17(9): 1066–83. https://doi.org/10.2174/1570180817666200304125520

40. Soofi AA, Awan A. Classification techniques in machine learning: applications and issues. Journal of Basic and Applied Sciences. 2017; 13: 459–65. https://doi.org/10.6000/1927-5129.2017.13.76

41. Obermeyer Z, Emanuel EJ. Predicting the future—big data, machine learning, and clinical medicine.The New England journal of medicine. 2016; 375(13): 1216-19. https://www.nejm.org/doi/full/10.1056/NEJMp1606181

42. Manoharan JS. Flawless detection of herbal plant leaf by machine learning classifier through two stage authentication procedure. Journal of Artificial Intelligence and Capsule Networks. 2021; Jun 22; 3(2): 125-39. https://doi.org/10.36548/jaicn.2021.2.005

43. Akter R, Hosen MI. CNN-based leaf image classification for Bangladeshi medicinal plant recognition. In: 2020 Emerging Technology in Computing, Communication and Electronics (ETCCE). IEEE, 2020; 1–6. https://doi.org/10.1109/ETCCE51779.2020.9350900

44. Lee, Han S, Chan CS, Wilkin P,Remagnino P. "Deep-plant: Plant identification with convolutional neural networks." In 2015 IEEE International Conference on Image Processing (ICIP), IEEE, 2015; 452-6. https://doi.org/10.1109/ICIP.2015.7350839

45. Yin H, Gu YH, Park C-J, Park J-H, Yoo SJ. Transfer learning-based search model for hot pepper diseases and pests. Agriculture 2020; 10(10): 1-16. https://doi.org/10.3390/agriculture10100439

46. Hassan SM, Maji AK, Jasiński M, Leonowicz Z, Jasińska E. Identification of plant-leaf diseases using CNN and transfer-learning approach. Electronics (Switzerland) 2021; 10(12): 1-19. https://doi.org/10.3390/electronics10121388

47. Pan SJ, Yang Q. A survey on transfer learning. IEEE Transactions on Knowledge and Data Engineering. 2010; 22(10): 1345–59. https://doi.org/10.1109/TKDE.2009.191

48. Mzoughi O, Yahiaoui I, Boujemaa N, Zagrouba E. Semantic shape models for leaf species identification. In: International Conference on Advanced Concepts for Intelligent Vision Systems. Springer, 2015; 661–71. https://doi.org/10.1007/978-3-319-25903-1_57

49. Ferentinos KP. Deep learning models for plant disease detection and diagnosis. Computerand Electronicsin Agriculture. 2018; 145: 311–8. https://doi.org/10.1016/j.compag.2018.01.009

50. Negi A, Pare A, Meenatchi R. Emerging techniques for adulterant authentication in spices and spice products. Food Control. 2021; 127: 108113. https://doi.org/10.1016/j.foodcont.2021.108113

51. Warhade VR, Dighe A. A review on quality control and standardization of herbals. Research Journal of Science and Technology 2022; 14(4): 247–252. http://dx.doi.org/10.52711/2349-2988.2022.00040

52. Kiani S, Ruth SM van, Raamsdonk LWD van, Minaei S. Hyperspectral imaging as a novel system for the authentication of spices: A nutmeg case study. Lwt. 2019; 104: 61–9. https://doi.org/10.1016/j.lwt.2019.01.045

53. Serrano N, Díaz-Cruz JM. Authentication of spices and herbs by chromatographic techniques. In: Chromatographic and Related Separation Techniques in Food Integrity and Authenticity: Volume B: Relevant Applications. World Scientific, 2021; 157–85. https://doi.org/10.1142/9781786349972_0006

54. Sasikumar B, Swetha VP, Parvathy VA, Sheeja TE. Advances in adulteration and authenticity testing of herbs and spices. In: Advances in Food Authenticity Testing. Elsevier, 2016; 585–624. https://doi.org/10.1016/B978-0-08-100220-9.00022-9

55. Oliveira MM, Cruz‐Tirado JP, Barbin DF. Nontargeted analytical methods as a powerful tool for the authentication of spices and herbs: A review. Comprehensive Reviews in Food Science and Food Safety. 2019; 18(3): 670–89. https://doi.org/10.1111/1541-4337.12436

56. Black C, Haughey SA, Chevallier OP, Galvin-King P, Elliott CT. A comprehensive strategy to detect the fraudulent adulteration of herbs: The oregano approach. Food Chemistry. 2016; 210: 551-7. https://doi.org/10.1016/j.foodchem.2016.05.004

57. Galvin-King P, Haughey SA, Elliott CT. Herb and spice fraud; The drivers, challenges and detection. Food Control. 2018; 88: 85–97. https://doi.org/10.1016/j.foodcont.2017.12.031

Received on 14.04.2024 Modified on 27.06.2024

Research J. Pharm. and Tech 2024; 17(10):4907-4914.

DOI: 10.52711/0974-360X.2024.00755