Prediction of banana quality attributes and ripeness classification using artificial neural network

Non-Destructive Banana Ripeness Detection Using Shallow and Deep Learning: A Systematic Review

“…A single banana is called a finger, and a grouping of attached fingers of 10 or more bananas is called a hand [ 3 ]. In this review, out of the 35 studies, a total of 27 [ 3 , 4 , 5 , 6 , 9 , 10 , 11 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 ] have used the fingers of bananas, 3 [ 37 , 38 , 39 ] studies have used the fingers by hand and Chen et al [ 40 ] used a bunch of bananas. Mohamedon et al and Rodrigues et al [ 41 , 42 ] used a mixed sample of bananas, Mendoza and Aguilera [ 43 ] used the combination of fingers by hand and a single batch of bananas, and whereas Saragih and Emanuel [ 44 ] also used the used the combination of fingers and fingers and hand bananas.…”

“…Apart from that, three [ 6 , 9 , 25 ] studies used the Egyptian species, and Sabilla et al [ 21 ] used local bananas from traditional markets in Indonesia (Pisang Emas, Pisang Kepok, Pisang Ambon, Pisang Raja Nangka, Pisang Santen, Pisang Susu, and Pisang Candi). Likewise, Saragih and Emanuel’s [ 44 ] study used the combination of two categories (Egyptian species and an unreported category), while the remaining 10 [ 11 , 18 , 24 , 26 , 30 , 34 , 37 , 38 , 41 , 42 ] studies did not report their banana categories.…”

“…A total of four of the 35 [ 4 , 17 , 18 , 35 ] studies have used the combination of Laser Light Backscattering Imaging (LLBI) with a wavelength distance and a Charge-Coupled Device (CCD). Another three [ 3 , 23 , 43 ] studies have used fluorescent lamps and digital cameras.…”

“…Medians were calculated from all the included augmentation and sensor studies for the quality assessment analysis. Out of all 30 quality assessment studies, 12 [ 6 , 9 , 10 , 17 , 18 , 19 , 20 , 25 , 35 , 38 , 41 , 44 ] have used 245 or more samples and 11 [ 4 , 5 , 22 , 23 , 27 , 30 , 33 , 37 , 40 , 43 ] studies have used values less than this amount. Among the four articles that have used a highly augmented dataset, Sabilla et al [ 21 ], as well as Marimuthu, Mohamed and Mansoor Roomi [ 39 ] have employed a median value of 2054 or more samples, and the other studies of Spaches [ 26 ] and Vaviya et al [ 34 ] used values less than this.…”

“…There are different stages used for the classification of banana ripeness. Among the 35 examined articles, 4 [ 5 , 22 , 36 , 43 ] examined seven stages, 4 [ 17 , 18 , 35 , 37 ] studies have used six stages, 6 [ 6 , 19 , 20 , 25 , 40 , 44 ] have used four stages, 12 [ 9 , 10 , 11 , 24 , 27 , 29 , 30 , 31 , 32 , 39 , 41 ] used three stages, and the remaining 3 [ 33 , 34 , 41 ] studies have used two stages. Lastly, Sabilla et al [ 21 ] used the seven categories of banana, and Iswari et al [ 28 ] employed three different types of fruits, namely, apple, banana and melon.…”