Plant diseases can cause a significant decrease in tea crop production. Early disease detection can help to minimize the loss. For tea plants, experts can identify the diseases by visual inspection on the leaves. However, providing experts to deal with disease identification may be very costly. The machine learning technology can be implemented to provide automatic plant disease detection. Currently, deep learning is state-of-the-art for object identification in computer vision. In this study, the researchers propose the Convolutional Neural Network (CNN) for tea disease detections. The researchers focus on the implementation of concatenated CNN, namely GoogleNet, Xception, and Inception-ResNet-v2, for this task. About 4727 images of tea leaves are collected, comprising of three types of diseases that commonly occur in Indonesia and a healthy class. The experimental results confirm the effectiveness of concatenated CNN for tea disease detections. The accuracy of 89.64% is achieved.
The capacity degradation of battery can occur due to continuously used as primary energy source equipment. An accurate prediction of battery remaining useful life (RUL) is necessary to avoid system functionality failure. This study proposes battery RUL prediction using data-driven method based on a hybrid deep model of Convolutional Neural Networks (CNN) and Long-Short Term Memory (LSTM). CNN and LSTM are used to extract features from multiple measurable data in parallel. CNN extracts features of multi-channel charging profiles, whereas LSTM extracts features of historical capacity data of discharging profiles which related to time dependency. An error index is compared between single model LSTM and hybrid model CNN-LSTM. The result indicates that the proposed hybrid model outperforms the single model by up to 37%-61% in case of mean absolute percentage error.
14 February 2021 in Journal of King Saud University Computer and Information Sciences
It is common to have various clones from cross-seedlings or unintended planting by the farmers in a tea plantation. Since each tea clone has distinctive features such as quality, resistance to diseases, etc., visual inspections are usually conducted on the plantations to segment areas with different tea clones within the plantation to produce crops with consistent quality. However, this would be costly and time-consuming. In this work, we apply machine learning and develop an application to recognize tea clones automatically. We propose a convolutional variational autoencoder-based feature learning algorithm to produce robust features against data distortions. There are two main advantages of using this algorithm for feature learning. First, there is no need to design complex handcrafted features for classifications, usually conducted in machine learning. Second, the resulting features are more robust when tested with data taken from unideal conditions. The proposed method is evaluated using the original and the distorted image. Our proposed method achieves the best performance of 0.83 (83%) for the original image test, 0.75 (75%) for the gaussian blur image test, and 0.78 (78%) for the median blur image test. This is a much more robust result than VGGNet16, a popular supervised deep convolutional neural network.
The rapid development of mobile devices has made human-computer interaction through voice increasingly popular and effective. This condition is made possible by the rapid growth of Automatic Speech Recognition (ASR) technologies. ASR can convert human speech signals into text and interpret it as a command for computer systems to perform. One of the remaining challenges for ASR is that the system’s performance can degrade significantly in noisy environments. This research aims to build a speech recognition system capable of recognizing human speech in clean and noisy conditions, enabling the system to recognize speech commands even in noisy conditions. Deep Neural Network (DNN) methods are the dominant methods for ASR. But, there is increasing interest in using a convolutional neural network (CNN) instead. In this study, we develop CNN-based architectures for robust ASR for speech commands. We explore various depths of CNN’s layers of CNN architecture to improve a robust speech recognition system. We also optimized the best model using early stopping and two types of optimizers, i.e., Adam and SGD (Stochastic gradient descent). Our experiment shows that CNN exhibited an accuracy of 90.64%, while the DNN model exhibited 86.74% accuracy in clean conditions. In noisy conditions, an increasing number of CNN layers improves ASR’s robustness. The CNN method achieves 77.38% accuracy in clean conditions and 87.59% in noisy conditions.
13 October 2020 in Jurnal Teknologi dan Sistem Komputer
Tea clone of Gambung series is a superior variety of tea that has high productivity and quality. Smallholder farmers usually plant these clones in the same areas. However, each clone has different productivity or quality, so it is difficult to predict the production quality in the same area. To uniform the variety of clones in an area, smallholder farmers still need experts to identify each plant because one and other clones share the same visual characteristics. We propose a tea clone identification system using deep CNN with skip connection methods, i.e., residual connections and densely connections, to tackle this problem. Our study shows that the proposed method is affected by the hyperparameter setting and the combining feature maps method. For the combining method, the concatenation method on a densely connected network shows better performance than the summation method on a residual connected network.
1 June 2023 in IAES International Journal of Artificial Intelligence (IJ-AI)
There are great interests in developing speech recognition using deep learning technologies due to their capability to model the complexity of pronunciations, syntax, and language rules of speech data better than the traditional hidden Markov model (HMM) do. But, the availability of large amount of data is necessary for deep learning-based speech recognition to be effective. While this is not a problem for mainstream languages such as English or Chinese, this is not the case for non-mainstream languages such as Indonesian. To overcome this limitation, we present deep features based on convolutional neural networks (CNN) for Indonesian large vocabulary continuous speech recognition in this paper. The CNN is trained discriminatively which is different from usual deep learning implementations where the networks are trained generatively. Our evaluations show that the proposed method on Indonesian speech data achieves 7.26% and 9.01% error reduction rates over the state-of-the-art deep belief networks-deep neural networks (DBN-DNN) for large vocabulary continuous speech recognition (LVCSR), with Mel frequency cepstral coefficients (MFCC) and filterbank (FBANK) used as features, respectively. An error reduction rate of 6.13% is achieved compared to CNN-DNN with generative training.