UMP Institutional Repository

The Effect of Image Input Transformation from Inertial Measurement Unit Data on the Classification of Skateboarding Tricks

Muhammad Amirul, Abdullah and Muhammad Ar Rahim, Ibrahim and Muhammad Nur Aiman, Shapiee and Muhammad Aizzat, Zakaria and Mohd Azraai, Mohd Razman and Musa, Rabiu Muazu and Anwar P. P., Abdul Majeed (2021) The Effect of Image Input Transformation from Inertial Measurement Unit Data on the Classification of Skateboarding Tricks. In: RiTA 2020: Proceedings of the 8th International Conference on Robot Intelligence Technology and Applications, 11-13 December 2020 , Virtual hosted by EUREKA Robotics Lab, Cardiff School of Technologies, Cardiff Metropolitan University. pp. 424-432.. ISBN 978-981-16-4803-8

[img] Pdf
The effect of image input transformation from inertial measurement unit data on the classification.pdf
Restricted to Repository staff only

Download (5MB) | Request a copy


This study aims to improve the classification of flat-ground tricks namely Ollie, Kickflip, Shove-it, Nollie and Frontside 180 through the identification of significant input image transformation on different transfer learning models. Six goofy skateboarders (23 years of age ± 5.0 years’ experience) executed five tricks for each type of trick repeatedly on a customized ORY skateboard (Inertial Measurement Unit (IMU) sensor fused) on a cemented ground. From the IMU data, six raw signals were extracted. The best input image transformation and transfer learning model were identified through two input image transformations synthesized, namely raw data (RAW) and Fast Fourier Transform (FFT), and six transfer learning models based on default arguments from the Keras library. The variation of the SVM models (via different hyperparameters) was evaluated both on input image transformation and on transfer learning model in classifying the skateboarding tricks. It was shown from the study that RAW input image on three transfer learning models (DenseNet121, InceptionResNetV2, and ResNet101) demonstrated the 100% accuracy on all train, train and validation dataset. It could be concluded that the proposed method is able to improve the classification of the skateboarding tricks well.

Item Type: Conference or Workshop Item (Lecture)
Additional Information: Part of the Lecture Notes in Mechanical Engineering book series (LNME)
Uncontrolled Keywords: Classification, Support, Vector, Machine, Machine learning, Skateboarding, Transfer learning
Subjects: T Technology > TJ Mechanical engineering and machinery
Faculty/Division: Faculty of Mechanical and Automotive Engineering Technology
Depositing User: Noorul Farina Arifin
Date Deposited: 26 Nov 2021 04:06
Last Modified: 26 Nov 2021 04:06
Download Statistic: View Download Statistics

Actions (login required)

View Item View Item