This study optimized the latest YOLOv5 framework, including its subset models, with training on different datasets that differed in image contrast and cloudiness to assess model performances based on quantitative metrics and image processing speed. The hyperparameter in the feature-extraction phase was configured based on the learning rate and momentum and further improved based on the adaptive moment estimation (ADAM) optimizer and the function reducing-learning-rate-on-plateau to optimize the model’s training scheme. The optimized YOLOv5s achieved a better performance, with a mean average precision of 98.6% and a high inference speed of 106 frames per second. The ADAM optimizer with a detailed learning rate (0.0001) and momentum (0.99) fine-tuning yielded a sufficient convergence rate (0.69% at 55th epoch) to assist YOLOv5s in attaining a more precise detection for underwater objects.
1 April 2023 in Bulletin of Electrical Engineering and Informatics
Deep learning has shown much promise in target identification in recent years, and it's becoming more popular in agriculture, where fig fruit detection and counting have become important. In this study, a systematic literature review (SLR) is utilised to evaluate a deep learning algorithm for detecting and counting fig fruits. The SLR is based on the widely used 'Reporting Standards for Systematic Evidence Synthetics' (ROSES) review process. The study starts by formulating the research questions, and the proposed SLR approach is critically discussed until the data abstraction and analysis process is completed. Following that, 33 relevant research involving the agriculture sector, fruit, were selected from many studies. IEEE, Scopus, and Web of Sciences are three databases to investigate. Due to the lack of fig fruit research, fruit and vegetable studies have been included because they use similar methods and processes. The SLR found that various deep learning algorithms can count fig fruit in the field. Furthermore, as most approaches obtained acceptable results, deep learning's performance is acceptable in F1-score and average precision (AP), higher than 80%. Moreover, improvements can be produced by enhancing the existing deep learning model with the personal dataset.
1 February 2024 in Bulletin of Electrical Engineering and Informatics
The process of identifying photos from a sketch has been explored by many researchers, and the performance of the identification process is almost perfect, particularly for viewed sketches. Suspect identification based on sketches is one of the applications in forensic science. To identify the suspect using these kinds of methods, a face sketch is required. Hence, the methods require skilled artists to sketch the suspect based on descriptions provided by eyewitnesses. However, the skills of these artists are different from one another, which results in different rendered sketches. Therefore, this work attempts to propose a new identification method based only on forensic face-written descriptions. To investigate the feasibility of the proposed method, this study has evaluated the performance of some text-to-photo generators on both viewed and forensic datasets using three different models of GAN which are SAGAN, DFGAN, and DCGAN. Then, the generated images are compared to the real photo contained within those datasets to evaluate how well the proposed method recognizes the faces. The results demonstrated that the recognition rate for the generated photos by the DCGAN models is better than the other two models which achieve a 38.3% recognition rate at rank-10 for mugshot identification.