1 February 2015 in International Journal of Electrical and Computer Engineering (IJECE)
Bus Route Planner applications will unfold their full potential when bus passengers are enabled to get information about the shortest path route, make a travel plan and get the correct buses in order to reduce the travel time. However, all these information are provided in text based and map view. It is difficult to understand them for the person who does not know place in the map. This paper describes the android base application of Augmented Reality (AR) that has feature to support the action of a bus user in an innovative and dynamic ways by putting additional information layer on smart phone camera screen and give the instruction assistant that leading the user way to the nearest bus stop. The experimental results show that, the overall functional of proposed application can be run well in various type of Android smart phone. When compared with similar bus traveling applications, the proposed application works more efficient.
1 March 2024 in IAES International Journal of Artificial Intelligence (IJ-AI)
Deep learning has shown outstanding performance in object detection tasks with unmanned aerial vehicles (UAVs), which involve the fine-tuning technique to improve performance by transferring features from pre-trained models to specific tasks. However, despite the immense popularity of fine-tuning, no works focused on to study of the precise fine-tuning effects of object detection tasks with UAVs. In this research, we conduct an experimental analysis of each existing fine-tuning strategy to answer which is the best procedure for transferring features with fine-tuning techniques. We also proposed a partial half fine-tuning strategy which we divided into two techniques: first half fine-tuning (First half F-T) and final half fine-tuning (Final half F-T). We use the VisDrone dataset for the training and validation process. Here we show that the partial half fine-tuning: Final half F-T can outperform other fine-tuning techniques and are also better than one of the state-of-the-art methods by a difference of 19.7% from the best results of previous studies.
1 February 2015 in International Journal of Electrical and Computer Engineering (IJECE)
Public transportation is important issue in Taiwan. Recently, mobile application named Bus Route Planning was developed to help the user to get information about public transportation using bus. But, this application often gave the user inaccurate bus information and this application has less attractive GUI. To overcome those 2 problems, it needed 2 kinds of solutions. First, a more accurate time prediction algorithm is needed to predict the arrival time of bus. Second, augmented reality technology can be used to make a GUI improvement. In this research, Automatic Data Collector system was proposed to give support for those 2 solutions at once. This proposed system has 3 main functionalities. First, data collector function to provide some data sets that can be further analyzed as an base of time prediction algorithm. Second, data updater functions to provide the most updated bus information for used in augmented reality system. Third, data management function to gave the system better functionality to supported those 2 related systems. This proposed Automatic Data Collector system was developed using batch data processing scenario and SQL native query in Java programming language. The result of testing shown this data processing scenario was very effective to made database manipulation especially for large-sized data.