Implementation of Transfer Learning Using VGG16 on Fruit Ripeness Detection

Abstract

In previous studies, researchers have determined the classification of fruit ripeness using the feature descriptor using color features (RGB, GSL, HSV, and L * a * b *). However, the performance from the experimental results obtained still yields results that are less than the maximum, viz the maximal accuracy is only 76%. Today, transfer learning techniques have been applied successfully in many real-world applications. For this reason, researchers propose transfer learning techniques using the VGG16 model. The proposed architecture uses VGG16 without the top layer. The top layer of the VGG16 replaced by adding a Multilayer Perceptron (MLP) block. The MLP block contains Flatten layer, a Dense layer, and Regularizes. The output of the MLP block uses the softmax activation function. There are three Regularizes that considered in the MLP block namely Dropout, Batch Normalization, and Regularizes kernels. The Regularizes selected are intended to reduce overfitting. The proposed architecture conducted on a fruit ripeness dataset that was created by researchers. Based on the experimental results found that the performance of the proposed architecture has better performance. Determination of the type of Regularizes is very influential on system performance. The best performance obtained on the MLP block that has Dropout 0.5 with increased accuracy reaching 18.42%. The Batch Normalization and the Regularizes kernels performance increased the accuracy amount of 10.52% and 2.63%, respectively. This study shows that the performance of deep learning using transfer learning always gets better performance than using machine learning with traditional feature extraction to determines fruit ripeness detection. This study gives also declaring that Dropout is the best technique to reduce overfitting in transfer learning.

Design a mobile application to detect tomato plant diseases based on deep learning

“…In this paper, a mobile application based on deep learning was created to classify tomato leaf diseases when a picture of the leaf plant was taken with a mobile camera. Transfer learning techniques of VGG16 [21], VGG19 [22], and MobileNet_v2 [23], [24] models were used in training the data in addition to using a proposed model for CNN [25]. The model that achieves the best result in the test dataset is converted into a tensorflow lite model (TFLM) open source deep learning framework to run pre-trained models on Android or iOS.…”

Knowledge Networks in Organic Fruit Production across Europe: A Survey Study

“…It is noteworthy that these "other subjects" were not always supported formally by the networks, thus pointing out that the need for knowledge about these topics is fulfiled by extemporary activities. Furthermore, it has been underlined that networks can be useful vehicles for knowledge transfer in new areas such as biodiversity management, climate change mitigation or quality certification [52,53], which have been emerging in recent years with regard to broad environment management. Interestingly, specialised working groups within the networks were established only in about half the networks, generally, those managed by research institutions or in networks active at national and international level (e.g., the institute of organic agriculture-FiBL or the Moroccan Interprofessional Federation of the Organic Sector-FIMABIO or the German-FOEKO), while seldom present in other kinds of networks (i.e., those referring to advisory bodies).…”

A Robust Fish Species Classification Framework: FRCNN-VGG16-SPPNet

“…Fine-tuning pre-trained models has been found to be an effective approach in adapting to new classi cation tasks. VGG16, a classic con-volutional neural network model, achieved outstanding performance in the ImageNet image classi cation competition and has been widely applied in various image classi ca-tion tasks [41][42][43][44][45][46], Therefore, VGG16 was selected as the backbone framework in this study. However, irrelevant features in images can affect classi cation accuracy.…”