Deenina Salleh
@utb.edu.bn
Lecturer, Creative Computing, School of Computing and Informatics
Universiti Teknologi Brunei
Scopus Publications
Scholar Citations
Scholar h-index
Scopus Publications
Muhd Amin Hj Fauzul and Noor Deenina Hj Mohd Salleh
Springer International Publishing
We develop a visual assistive system to aid the visually impaired users for safe and convenient navigation in indoor and outdoor environments. The system has two main components: a mobile app and an obstacle sensor. The mobile app makes use of the microelectromechanical sensors inside the smartphone, location services, and Google Maps to provide audio cues directions for the user. The obstacle sensor employs ultrasonic sensors to detect obstacles and provide haptic feedback. The obstacle avoidance device can be attached to a traditional probing cane, and the device vibrates the probing cane handle at different intensity according to the nature of obstacles. Spatial sound cues are generated based on the spatial distance and direction of the current location to the desired destination. We discuss the system design and report the testing to evaluate the device effectiveness for obstacle distance feedback as well as the effectiveness of the spatial sound cues for navigation feedback.
Somnuk Phon-Amnuaisuk, Noor Deenina Hj Mohd Salleh, and Siew-Leing Woo
Springer International Publishing
Applying computational intelligence techniques to create generative models of digits or alphabets has received somewhat little attention as compared to classification task. It is also more challenging to create a generative model that could successfully capture styles and detailed characteristics of symbols. In this paper, we describe the application of the Long Short-Term Memory (LSTM) model trained using a supervised learning approach for generating a variety of the letter A. LSTM is a recurrent neural network with a strong salient feature in its ability to handle long range dependencies, hence, it is a popular choice for building intelligent applications for speech recognition, conversation agent and other problems in time series domains. To formulate the problem as a generative task, all the pixels in a 2D image representing an alphabet (i.e., the letter A in this study) are flattened into a long vector to train the LSTM model. We have shown that LSTM has successfully learned to generate new letters A showing many coherent stylistic features with the original letters from the training sets.
Noor Deenina Hj Mohd Salleh and Somnuk Phon-Amnuaisuk
Springer International Publishing