Research Publications
Permanent URI for this communityhttps://rda.sliit.lk/handle/123456789/4194
This main community comprises five sub-communities, each representing the academic contribution made by SLIIT-affiliated personnel.
Browse
2 results
Search Results
Publication Open Access Karate Kata Scoring and Performance Evaluation Using Video Analysis and Deep Learning(Faculty of Engineering, 2025-09-09) Liyanage, H.L.S.S.; Deshpriya, H.M.S.D.; Kavindya,M.T.; Ranasinghe, R.M.L.D; Herath, H.M.D.P.; Aththanagoda, A.K.N.L.Karate kata, a fundamental element of traditional martial arts, consists of choreographed sequences of stances, strikes, and blocks performed against imaginary opponents, requiring precise posture, timing, and technique. However, self-practice often lacks objective feedback, leading to unnoticed errors and uneven skill development, and even competition scoring can be influenced by subjective judgment. To address this, we present a vision-based scoring system that evaluates kata from structured video recordings without wearable sensors, providing accurate, consistent scores to help practitioners track progress. The system processes videos through a multi-stage pipeline: frames are extracted at 10 frames per second using OpenCV and segmented with YOLOv8-seg to remove background clutter. Pose estimation is performed with MediaPipe, generating 99 normalized 3D keypoints, while hip-centered normalization and pelvic-width scaling ensure consistency across different body sizes. A modified ResNet50 classifies frames into nine fundamental stances with high accuracy and near real-time processing (~150 ms per frame at 1080p), and a custom Convolutional Neural Network evaluates the angular accuracy of 14 critical joints using trigonometric analysis against reference poses. Each kata is decomposed into 19 key positions, with a stance marked correct only if both classification and joint angles (≤ ±2.5° deviation) meet the threshold, and final scores are calculated on a 5.0–10.0 scale to align with traditional judging. Experimental results show that ResNet50 achieves 95.97% accuracy, while the CNN reaches 96.11%, demonstrating that this approach offers a low-cost, accessible, and consistent scoring tool. It supports remote training, reduces judging bias, and provides practitioners with reliable feedback to improve their kata performance.publications.listelement.badge Open Access Automated Flood Impact Detection from Remote Sensing Data using Transfer Learning: A New Zealand Case Study(Sri Lanka Institute of Information Technology, 2026-05-21) Pathirana, N; Ranasinghe, M; Prasanna, RProper flood damage assessment is required for rapid emergency response and recovery planning. Highresolution drone imagery provides greater visual detail compared to satellite imagery and can support more accurate damage assessment after flood events. This work-in-progress study investigates the use of transfer learning with a pre-trained ResNet50 convolutional neural network to classify post-flood drone images. A manually annotated dataset of drone images representing damaged and non-damaged areas was used to finetune the pre-trained model and evaluate its performance using several metrics. The training process employed hyperparameter tuning to select the best model which achieved an accuracy of approximately 87%. These preliminary results demonstrate the potential of transfer learning for flood damage classification using a limited drone image dataset. Future work will focus on comparing the performance of additional deep learning models, incorporating satellite imagery and extending the approach from image-level classification to object-level damage detection.