Automation testing is essential to carry out functional testing quickly and precisely. Software testing is beneficial for testers doing many testing processes according to the existing scenarios. So, there is an urgen...
详细信息
- An online learning system is a learning approach that typically uses a one fit for all approach in which all student abilities are considered the same so that the provision of learning materials provided is the same...
详细信息
Vision impairment, often caused by preventable ocular diseases can be challenging to diagnose accurately and prone to human error. Automation using technology, particularly deep learning, offers a promising solution t...
详细信息
Vision impairment, often caused by preventable ocular diseases can be challenging to diagnose accurately and prone to human error. Automation using technology, particularly deep learning, offers a promising solution to aid in accurate and efficient disease detection. This study explores the use of different CNN models specifically VGG-16, VGG-19, ResNet-50, and ResNet-152v2, for detecting ocular diseases. Simple fine-tuning is applied to these models, and their performance is compared to identify the most effective model. The purpose is to show how different models contribute to establishing reliable illness detection systems. The results reveal that most of these models perform well with even minimal fine-tuning. Among the models, ResNet-152v2 achieved the highest training accuracy of 90.36% demonstrating its capacity to learn from the training data. In contrast, ResNet-50 offered a more balanced performance with marginally lower accuracy, making it a robust choice for general application.
The industry is rapidly transitioning from the 4.0 era to the 5.0 era, prompting renewed interest among scholars in scheduling problems. They allow operations to process and assemble various components simultaneously....
详细信息
Neural style transfer is a technique to transfer a style of an artistic image to another photorealistic image. The early development of this technique is exploiting the intermediate output of a deep learning model and...
详细信息
The design of a hybrid Hematology I course (prepandemic) was adopted for Hematology II and facilitated our conversion of Hematology II into a fully remote course by fall 2020, after the university went into remote ins...
Membership function (MF) in process of fuzzy logic is very meaningful. It depicts the core of model. It can be adopted from the expert judgment and also coming from the configuration of data behavior. The study is an ...
详细信息
Social media is an online media that functions as a platform for users to participate, share, create, and exchange information through various forums and social networks. The rapid increase in social media activity ca...
详细信息
With the increase of interest in Facial Expression Recognition (FER) in the past few decades. Several challenges surfaced with the invention of many different FER models which are often based on Convolution Neural Net...
详细信息
With the increase of interest in Facial Expression Recognition (FER) in the past few decades. Several challenges surfaced with the invention of many different FER models which are often based on Convolution Neural Network (CNN) architectures. Recently, an attention-based transformer model has been presented to address FER. One of the major issues with Transformers is the need for a large data quantity for training. Therefore, in this paper, we propose to learn how to fine-tune a vision transformer-based (ViT) model using a limited dataset. We will be using the JAFFE Dataset, which consists of only 213 images containing seven different emotions. The proposed method is evaluated using several fine-tuning methods, such as adding dropout, data augmentation, and layer freezing. We compared the models implemented with 5% dropout regularization, augmented dataset (up to 5000 images), and freezing the initial model's layers, fine-tuning around a fourth of the last layers. The best model was achieved by fine-tuning ViT L-16 with 96.06% accuracy, trained with 5% dropout in the augmented dataset, and freezing the initial 21st layers. We also compared our model to the other previous work model and the results showed that our model reached the state-of-the art for the JAFFE dataset.
Competitive programming (CP) is a mind sports activity where people solve problems using command-line computerprograms to provide correct output for the given test cases. Competitors need to practice problem-solving ...
详细信息
Competitive programming (CP) is a mind sports activity where people solve problems using command-line computerprograms to provide correct output for the given test cases. Competitors need to practice problem-solving and mathematics as well as study algorithms and data structures to perform well in CP. This study aims to provide an original way to perform a trend analysis in CP, distinguishing topics frequently used in CP contests. To fulfill our goal, we create topic models based on previous topic modeling works to do natural language processing tasks using Latent Dirichlet Allocation (LDA) and Biterm Topic Model (BTM). For our dataset, we constructed a corpus from Codeforces blog posts, a popular website for competitive programmers, by extracting its content and user comments. Our results indicate that BTM is powerful enough to do trend analysis in CP. The trend analysis recognized that dynamic programming and complexity analysis have been the most prominent topics for the last ten years. Data structures and string algorithms are runners-up that may have potential trends in the future. This study opens up further research on other methods to perform trend analysis using better topic models and corpora.
暂无评论