Browsing by Author "Liyanapathirana, C"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    item: Conference-Full-text
    Infinity yoga tutor: yoga posture detection and correction system
    (Faculty of Information Technology, University of Moratuwa., 2020-12) Rishan, F; De Silva, B; Alawathugoda, S; Nijabdeen, S; Rupasinghe, L; Liyanapathirana, C; Karunananda, AS; Talagala, PD
    Popularity of yoga is increasing daily. The reason for this is the physical, mental and spiritual benefits that could be obtained by practicing yoga. Many are following this trend and practicing yoga without the training of an expert practitioner. However, following yoga in an improper way or without a proper guidance will lead to bad health issues such as strokes, nerve damage etc. So, following proper yoga postures is an important factor to be considered. In this proposed system, the system is able to identify poses performed by the user and also guide the user visually. This process is required to be completed in real-time in order to be more interactive with the user. In this paper, the yoga posture detection was done in a vision-based approach. The Infinity Yoga Tutor application is able to capture user movements using the mobile camera, which is then streamed at a resolution of 1280 × 720 at 30 frames per second to the detection system. The system consists of two main modules, a pose estimation module which uses OpenPose to identify 25 keypoints in the human body, using the BODY_25 dataset, and a pose detection module which consists of a Deep Learning model, that uses time-distributed Convolutional Neural Networks, Long Short Term Memory and SoftMax regression in order to analyze and predict user pose or asana using a sequence of frames. This module was trained to classify 6 different asanas and the selected model which uses OpenPose for pose estimation has an accuracy of 99.91%. Finally, the system notifies the users on their performance visually in the user interface of the Mobile application.
  • Thumbnail Image
    item: Conference-Full-text
    Vision based intelligent shelf-management system
    (Faculty of Information Technology, University of Moratuwa., 2021-12) Priyanwada, HAM; Madhushan, KADD; Liyanapathirana, C; Rupasinghe, L; Ganegoda, GU; Mahadewa, KT
    Currently supermarkets are more popular, and the local stores are leaving the competition. when people go to supermarkets, they find various items stocked on seemingly unlimited shelves. Supermarket shelves needed to be filled with the items accordingly. The most common problems in the supermarkets are identifying the empty shelves, on-shelf availability, and future sales. The labors cannot always track the empty shelves and on shelf availability levels due to their workloads. Moreover, it is a time-consuming method for the labors which can affect the customer satisfaction and business profit. Every month, supermarkets buy the required number of products from related manufacturing companies by analyzing the previously purchased products and their sales. This is usually done manually by managing excel sheets which is also time consuming and not reliable. Especially during the seasonal times or pandemic situations they cannot use the manual method which must also be done as fast as possible. Therefore, this system can be used to assist in empty shelf detection, percentage of on-shelf availability and in the prediction of future sales. The implementation of on-shelves percentage detection service is done using machine learning. Machine learning processes are carried out for implementing the necessary functionalities and algorithms. Initially, the camera captures clear and real time images regularly. Then the system processes and detects the image similar to the threshold percentage or detect the empty shelves. When the system detects the threshold percentage or empty shelves, the system will provide an alert to the labors. The Implementation of the predicting the future supply and demands is done using time series analysis using several existing machine learning algorithms by utilizing historical data. In this research the prediction of future sales and demand in the supermarkets is done by considering the customers' behavior, the variety of product groups they buy and seasonal changes. These predictions are made on the assumption of a constant per capital supply of products and demand in our system.