1 October 2024 in Indonesian Journal of Electrical Engineering and Computer Science
A leading cause of mortality globally, skin cancer is deadly. Early skin cancer diagnosis reduces mortality. Visual inspection is the main skin cancer diagnosis tool; however, it is imprecise. Researchers propose deep-learning techniques to assist physicians identify skin tumors fast and correctly. Deep convolutional neural networks (CNNs) can identify distinct objects in complex tasks. We train a CNN on photos with merely pixels and illness labels to classify skin lesions. We train on HAM-10000 using a CNN. On the HAM10000 dataset, the suggested model scored 95.23% efficiency, 95.30% sensitivity, and 95.91% specificity.
1 December 2024 in Indonesian Journal of Electrical Engineering and Computer Science
In blood smear images, there are difficulties in diagnosing blood cancer diseases like leukemia and lymphoma because of their various forms that appear in the human body. In this paper, a method for automatic detection of blood cancer is suggested that uses the EfficientNet-B3 architecture along with transfer learning techniques to improve accuracy and efficiency. We first fine-tuned the EfficientNet-B3 model, which was pre-trained on a large dataset consisting of annotated blood smear images, to capture pertinent features linked with blood malignant cells. To expedite the training process and adapt the model to our task, we use transfer learning. The proposed approach’s results from our experiments show that it outperforms traditional deep learning models and state-of-the-art methods in blood cancer detection. Additionally, with high precision and recall rates, this model also detects different types of blood cancers with robustness in its performance since its accuracy is over 99%. This means that when used together with the EfficientNet-B3 architecture, transfer learning can help the developed methods generalize among different types of blood cancers and conditions.
5 September 2019 in IEEE Access
In this paper, we deal with one of the current challenges in process mining enhancement: the prediction of remaining times in business processes. Accurate predictions of the remaining time, defined as the required time for an instance process to finish, are critical in many systems for organisations being able to establish a priori requirements, for optimal management of resources or for improving the quality of the services organisations provide. Our approach consists of i) extracting and assessing a number of features on the business logs, that provide a structural characterisation of the traces; ii) extending the well-known annotated transition system (ATS) model to include these features; iii) proposing a partitioning strategy for the lists of features associated to each state in the extended ATS; and iv) applying a linear regression technique to each partition for predicting the remaining time of new traces. Extensive experimentation using eight attributes and ten real-life datasets show that the proposed approach outperforms in terms of mean absolute error and accuracy all the other approaches in state of the art, which includes ATS-based, non-ATS based as well as Deep Learning-based approaches.
1 May 2025 in Indonesian Journal of Electrical Engineering and Computer Science
In the last few years, deep learning has exhibited its efficacy and capacity in the field of computer vision owing to its exceptional precision and widespread acceptance. The primary objective of this study is to investigate an improved approach for segmentation in the context of various fruit categories. Despite the utilization of deep learning, the current segmentation techniques for various fruit items exhibit subpar performance. The proposed enhanced multiple fruit segmentation algorithm has the following main steps: 1) modifying the size of the filter, 2) the process of optimizing the ResNet-101 block involves selecting the most suitable count of repetitions. The multiple fruit dataset is split 80% in the training stage and 20% in the testing stage. These images were utilized to train a deep learning (DL) based algorithm, which aims to identify multiple fruit items within images accurately. The proposed algorithm has a lower training time compared to the other algorithms. The thresholds exhibit greater values compared to the thresholds of state-of-the-art algorithms.