Table 1 Analysis of the previous work.

Khan et al. (2022)⁶

Data of 20 food items

Bi-LSTM, iHearken

Accuracy = 97.42%

The work could be extended by incorporating more advanced techniques for the classification of food items

Precision = 96.80%

Recall = 98%

F1 score = 97.51%

Kojima et al. (2016)⁷

Data of six fruits and vegetables

KNN

Accuracy = 83%

Limited dataset

SVM

Accuracy = 95%

CNN

Accuracy = 89%

Vijayakumari et al. (2022)⁸

101 different food products

EfficientNetB0

Accuracy = 80%

The model should in future be applied to both image as well as text data

Gao et al. (2016)⁹

Data collected from 28 individuals

SVM

Accuracy = 95%

No diversity had been seen in the dataset

Uchiyama et al. (2021)¹⁰

Data of food ASMR video collected from YouTube

Spectrogram, inverse STFT, Griffin-Lim algorithm

Perceptual evaluation of speech quality (PESQ) = 1.27

The algorithm could be applied to real time data

Päßler and Fischer (2016)¹¹

68,094 chewing sounds

Biomedical signal processing

Precision = 80%

The system needed an optimization to enhance its performance

Amft et al. (2009)¹²

Data taken from eight participants

Pattern Recognition Procedure

Precision = 70%

The model could be applied only for the solid foods

Recall = 80%

Amft et al. (2005)¹³

Four various types of food

Hearing aids, Headsets

Accuracy = 99%

Limited dataset