Classification of first quality fancy cashew kernels using four deep convolutional neural network models

Abstract

In this study, we proposed deep convolutional neural networks (DCNNs) combined with image processing to classify cashew kernels in five categories based on the adulteration of first-class fancy whole cashew kernels with butts and pieces. Four DCNN models, including Inception-V3, ResNet50, VGG-16, and a custom model were implemented, and their performances were compared using model evaluators, such as sensitivity, specificity, precision, accuracy, and F1-score. Overall, all the models showed a high performance with a minimum accuracy of 95.1% and the training and validation data curves demonstrated a good fitting of the models. Although all the models demonstrated promising potential for cashew classification, Inception-V3 and ResNet50 neural networks delivered the most promising outcome with the highest accuracies (each 98.4%) and F1-scores (each 96%). Custom-built model showed the least accuracy (95.1%) and F1-score (87.9%). The findings of current work indicate that the developed DCNN models are capable of achieving automatic, fast, and accurate cashew classification. Practical application Currently, the sorting and grading of cashew kernel are mostly manual and timeconsuming process and the deep convolutional neural networks (DCNN) or Deep Learning implemented in this study can facilitate a speedy and automated cashew classification in cashew industry worldwide. 1 | INTRODUCTION 1.1 | General overview of cashew nuts and their geographical production/distribution Cashew (Anacardium occidentale L.) nuts belong to a tropical evergreen tree from the Anacardiaceae family of plants, which also includes mango, pistachio, poison oak, and poison ivy. They contain about 35-45% seeds and 55-65% shells by weight (Mgaya et al., 2019). Cashew nuts are kidney-shaped nuts which are rich in unsaturated fatty acids, high protein, and low levels of saturated fats and sugars (Rao, 1998). Cashew nuts are an important ingredient in snack and confectionary industries and are generally consumed in raw, roasted, salted or flavored forms (Papadenetriou, 1998). About 60% of the cashew kernels are consumed in the form of snacks and remaining 40% are included in other food preparations, mainly confectionary (Azam-Ali & Judge, 2001). In 2017, the worldwide production of raw cashew nuts was about 4 million metric tons and it is expected that cashew nuts will take over 29% of total nut market by 2021 (Yun, 2018). The main cashew production regions are Vietnam, India, Africa, and Philippines. Vietnam is considered the largest cashew exporter, followed by India, Netherlands, and African countries (Yun, 2018). Cashew nuts have an established market in the US with a wide range of uses and is considered the largest importer in the world accounting for about 35% of the total import (Azam-Ali & Judge, 2001).

Traditional Machine Learning-Based Classification of Cashew Kernels Using Colour Features

“…A novel intelligent model extracted 24 colour and 16 morphological features of the cashew kernel and used a Multilayer Perceptron ANN to recognize and classify white wholes into different grades using a Backpropagation learning algorithm and attained a classification of 88.93% (Ganganagowdar and Siddaramappa, 2016). Vidyarthi et al (2020) applied four different deep Convolutional Neural Network models, which are VGG-16, ResNet50, Inception-V3 and a custom model to classify cashew kernels into five categories. The overall minimum accuracy of all the models was 95.1%.…”

A Novel Approach to Cashew Nut Detection in Packaging and Quality Inspection Lines

“…However, this method can only classify wholes and others (scorched whole, splits, butts, pieces). In [24], four deep CNNs combined with image processing were used to classify cashew kernels into five categories based on their adulteration with butts and pieces. Therefore, the above studies demonstrate the effectiveness of machine vision in the classification and grading of cashew nuts.…”

Öznitelik Seviyesinde Füzyon Yaklaşımının Kuruyemiş Tür Sınıflandırılmasında Performans Değerlendirmesi

“…yaptıkları çalışmada derin Evrişimli Sinir Ağlarını (ESA) kullanarak kaju çekirdeklerini, beş farklı sınıf etiketi üzerinden sınıflandırmayı önermişlerdir. Kendi önerdikleri özel ESA modelleri dışında, üç farklı transfer öğrenme ağını kullanarak elde ettikleri başarı çıktılarına bakıldığında, Inceptionv3 ve ResNet50 ağlarının %98,4'lük sınıflandırma doğruluğuna ulaştığı raporlanmıştır [2].Dheir, I. M. vd. ise yaptıkları çalışmada geniş bir kuruyemiş yelpazesini barındıran veri kümesi ile çalışmışlardır.…”