Browsing by Author "Mohottala, S"

Now showing 1 - 6 of 6

Open Access
2D Pose Estimation based Child Action Recognition
(Institute of Electrical and Electronics Engineers Inc., 2022-11) Mohottala, S; Abeygunawardana, S; Samarasinghe, P; Kasthurirathna, D; Abhayaratne, C
We present a graph convolutional network with 2D pose estimation for the first time on child action recognition task achieving on par results with LRCN on a benchmark dataset containing unconstrained environment based videos.
Embargo
Assessing the Efficacy of Machine Learning Algorithms in Predicting Critical Properties of Gold Nanoparticles for Pharmaceutical Applications
(Springer Nature Link, 2025-07-08) Fernando, H; Mohottala, S; Jayanetti, M; Thambiliyagodage, C
Au nanoparticles are increasingly used in pharmaceuticals, but their synthesis is costly and time-intensive. Machine Learning can help optimize this process. In this research, eight distinct Machine Learning models were implemented and optimized on a dataset comprising 3000 records of gold nanoparticles. The performance of these models was assessed using four accuracy metrics and the time required for training and inference. The results are promising, with all seven models demonstrating high accuracy and low time requirements. Notably, the XGBoost and Artificial Neural Network models exhibited exceptional performance, with Mean Squared Error values of 0.0235 and 0.0098, Mean Absolute Error values of 0.1021 and 0.0674, Mean Absolute Percentage Deviation values of 0.4945 and 0.3590, R2 scores of 0.9995 and 0.9998, and inference times of 0.0029 and 0.4299 s, respectively. The Explainable Artificial Intelligence analysis of the resulting models revealed some interesting insights into how the models make the predictions and what factors heavily contribute to the nanoparticle AVG_R, allowing chemists to optimize the synthesis for gold nanoparticles better. The key contributions of the research include the design and development of eight Machine Learning models using industry-standard frameworks, the training, tuning, and evaluation of these eight models using five different metrics, and further assessment of these trained models using Explainable Artificial Intelligence. The findings indicate a substantial potential for applying neural networks in the design phase of nanoparticle synthesis, which could lead to significant reductions in both the time and cost required for synthesizing Au nanoparticles for pharmaceutical applications.
Embargo
Assessing the Efficacy of Machine Learning Algorithms in Predicting Critical Properties of Gold Nanoparticles for Pharmaceutical Applications
(Springer, 2025-07-08) Fernando, H; Mohottala, S; Jayanetti, M; Thambiliyagodage, C
Au nanoparticles are increasingly used in pharmaceuticals, but their synthesis is costly and time-intensive. Machine Learning can help optimize this process. In this research, eight distinct Machine Learning models were implemented and optimized on a dataset comprising 3000 records of gold nanoparticles. The performance of these models was assessed using four accuracy metrics and the time required for training and inference. The results are promising, with all seven models demonstrating high accuracy and low time requirements. Notably, the XGBoost and Artificial Neural Network models exhibited exceptional performance, with Mean Squared Error values of 0.0235 and 0.0098, Mean Absolute Error values of 0.1021 and 0.0674, Mean Absolute Percentage Deviation values of 0.4945 and 0.3590, R2 scores of 0.9995 and 0.9998, and inference times of 0.0029 and 0.4299 s, respectively. The Explainable Artificial Intelligence analysis of the resulting models revealed some interesting insights into how the models make the predictions and what factors heavily contribute to the nanoparticle AVG_R, allowing chemists to optimize the synthesis for gold nanoparticles better. The key contributions of the research include the design and development of eight Machine Learning models using industry-standard frameworks, the training, tuning, and evaluation of these eight models using five different metrics, and further assessment of these trained models using Explainable Artificial Intelligence. The findings indicate a substantial potential for applying neural networks in the design phase of nanoparticle synthesis, which could lead to significant reductions in both the time and cost required for synthesizing Au nanoparticles for pharmaceutical applications.
Embargo
Graph Neural Network based Child Activity Recognition
(IEEE, 2022-08-25) Mohottala, S; Samarasinghe, P; Kasthurirathna, D; Abhayaratne, C
This paper presents an implementation on child activity recognition (CAR) with a graph convolution network (GCN) based deep learning model since prior implementations in this domain have been dominated by CNN, LSTM and other methods despite the superior performance of GCN. To the best of our knowledge, we are the first to use a GCN model in child activity recognition domain. In overcoming the challenges of having small size publicly available child action datasets, several learning methods such as feature extraction, fine-tuning and curriculum learning were implemented to improve the model performance. Inspired by the contradicting claims made on the use of transfer learning in CAR, we conducted a detailed implementation and analysis on transfer learning together with a study on negative transfer learning effect on CAR as it hasn’t been addressed previously. As the principal contribution, we were able to develop a ST-GCN based CAR model which, despite the small size of the dataset, obtained around 50% accuracy on vanilla implementations. With feature extraction and fine tuning methods, accuracy was improved by 20%-30% with the highest accuracy being 82.24%. Furthermore, the results provided on activity datasets empirically demonstrate that with careful selection of pre-train model datasets through methods such as curriculum learning could enhance the accuracy levels. Finally, we provide preliminary evidence on possible frame rate effect on the accuracy of CAR models, a direction future research can explore.
Embargo
Model Development for Child Developmental Milestone Assessment
(IEEE, 2022-12-09) Mudannayake, M; Kumari, G; Abeygunawardana, S; Amaranayake, N; Samarasinghe, P; Mohottala, S; Wijethunga, S
This paper presents the implementation of models for assessing the developmental milestone of children below age five on physical, cognitive, social, and emotional factors, which is a crucial aspect of human development. To the best of our knowledge, we are the first to design and evaluate models assessing the developmental delays of Sri Lankan children. The primary goal of this study is to analyze the ability of children to reach the relevant milestones in their childhood using video recordings and parents’ feedback. Out of the different models we experimented with, we selected the best models in our final evaluation. As the principal contribution, we developed a model to decide whether the child has typical developmental growth or otherwise using parents’ feedback and obtained 92.31% accuracy. Furthermore, we achieved 92.76% for the social and emotion detection model and 88.44% accuracy for the child action recognition model using video-based datasets. In the future implementation, the derived models will be integrated to build a mobile application.
Embargo
Spatio-temporal graph neural network based child action recognition using data-efficient methods: A systematic analysis
(Elsevier Inc, 2025-06-03) Mohottala, S; Gawesha, A; Kasthurirathna, D; Samarasinghe, P; Abhayaratne, C
This paper presents implementations on child activity recognition (CAR) using spatial–temporal graph neural network (ST-GNN)-based deep learning models with the skeleton modality. Prior implementations in this domain have predominantly utilized CNN, LSTM, and other methods, despite the superior performance potential of graph neural networks. To the best of our knowledge, this study is the first to use an ST-GNN model for child activity recognition employing both in-the-lab, in-the-wild, and in-the-deployment skeleton data. To overcome the challenges posed by small publicly available child action datasets, transfer learning methods such as feature extraction and fine-tuning were applied to enhance model performance. As a principal contribution, we developed an ST-GNN-based skeleton modality model that, despite using a relatively small child action dataset, achieved superior performance (94.81%) compared to implementations trained on a significantly larger (x10) adult action dataset (90.6%) for a similar subset of actions. With ST-GCN-based feature extraction and fine-tuning methods, accuracy improved by 10%–40% compared to vanilla implementations, achieving a maximum accuracy of 94.81%. Additionally, implementations with other ST-GNN models demonstrated further accuracy improvements of 15%–45% over the ST-GCN baseline. The results on activity datasets empirically demonstrate that class diversity, dataset size, and careful selection of pre-training datasets significantly enhance accuracy. In-the-wild and in-the-deployment implementations confirm the real-world applicability of above approaches, with the ST-GNN model achieving 11 FPS on streaming data. Finally, preliminary evidence on the impact of graph expressivity and graph rewiring on accuracy of small dataset-based models is provided, outlining potential directions for future research. The codes are available at https://github.com/sankamohotttala/ST_GNN_HAR_DEML.