The Use of Different Image Recognition Techniques in Food Safety: A Study

Table 3

Comparative study of existing image recognition techniques for food safety.

Reference	Preprocessing techniques	No. of categories	Training	Color space	Evaluation criteria
Danti et al. [35]	Cropping and resizing	10	BPNN; Classifier	HSI	Accuracy; 96.40%
Rocha et al. [36]	K-means with two clusters	15	Multiclass; SVM	HSI	Average error; 97%
Suresha et al. [37]	Watershed; Segmentation	8	Decision tree classifier	RGB	Accuracy; 95%
Faria et al. [38]	K-means with two clusters	15	Classifier fusion	HSI	Accuracy; 98.8%±0.9
Kanakaraddi et al. [39]	Median filtering	4	Decision tree	RGB	Disease; Severity
Pydipati et al. [40]	Edge detection	4	Generalized; Squared distance	HSI	Accuracy; >95%
Pujari et al. [41]	K-means	2	BPNN; Classifier	RGB	Accuracy; 84.65%
Kim et al. [42]	ROI cropping	6	Discriminant; Analysis	HSI	Accuracy; 96%