Department of Computer Science and Engineering

Permanent URI for this collectionhttp://dspace.iiuc.ac.bd/handle/88203/1527

Browse

Start date: 2020End date: 2022

Search Results

Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    HActivityNet: A Deep Convolutional Neural Network for Human Activity Recognition
    (EMITTER International Journal of Engineering Technology ‐ Published by EEPIS, 2021-12) Khaliluzzaman, Md.; Sayem, Abu Bakar Siddiq; Misbah, Lutful Kader
    Human Activity Recognition (HAR), a vast area of a computer vision research, has gained standings in recent years due to its applications in various fields. As human activity has diversification in action, interaction, and it embraces a large amount of data and powerful computational resources, it is very difficult to recognize human activities from an image. In order to solve the computational cost and vanishing gradient problem, in this work, we have proposed a revised simple convolutional neural network (CNN) model named Human Activity Recognition Network (HActivityNet) that is automatically extract and learn features and recognize activities in a rapid, precise and consistent manner. To solve the problem of imbalanced positive and negative data, we have created two datasets, one is HARDataset1 dataset which is created by extracted image frames from KTH dataset, and another one is HARDataset2 dataset prepared from activity video frames performed by us. The comprehensive experiment shows that our model performs better with respect to the present state of the art models. The proposed model attains an accuracy of 99.5% on HARDatase1 and almost 100% on HARDataset2 dataset. The proposed model also performed well on real data.
  • No Thumbnail Available
    Item
    Face and Hand Gesture Recognition Based Person Identification System using Convolutional Neural Network
    (International Journal of Intelligent Systems and Applications in Engineering, 2022-02-07) Kabisha, Mysha Sarin; Rahim, Kazi Anisa; Khaliluzzaman, Md.; Khan, Shahidul Islam
    Person identification system is now become the most hyped system for security purpose. It’s also gaining a lot of attention in the field of computer vision. For verification of human, facial recognition and hand gesture recognition are the most common topics of research. In the current days, various researchers focused on facial and hand gesture recognition using various shallow techniques and Deep Convolutional Neural Network (DCNN). However, using one feature of human for person identification is the most researched topic till now. In this paper, we proposed a Convolutional Neural Network (CNN) based system which will identify a person using two traits i.e., face and hand gesture of number sign of that person. For feature extraction and recognition Neural Network have shown immense good result. This proposed system works on two models, one is a VGG16 architecture model for face recognition and another model is for hand gesture which is based on simple CNN with two convolutional layers. With two customized dataset our face model gained 98.00% accuracy and hand gesture (number sign) model gained an accuracy of 98.33%.
  • No Thumbnail Available
    Item
    Analytical justification of vanishing point problem in the case of stairways recognition
    (Journal of King Saud University Computer and Information Sciences, 2021) Khaliluzzaman, Md.
    Stair region detection and recognition from a stair candidate image is a challenging work in the computer vision research area. In the last few decades, researchers use many recognition systems to recognize and verify the stair region from other analogous objects. However, all the verification systems such as vanish ing point (VP) do not achieve the desired result for various reasons. In this regard, a method is proposed in this paper to investigate the vanishing point’s problem arising in the case of stair region verification based on the three basic criteria, i.e. focal angle of the camera, height of the camera from the ground, and distance of the camera from the stair image. For that, primarily, the stair region is extracted by uti lizing the geometrical features of a stair. The detected stair candidate region is verified through the y coordinate value of the vertical VP, i.e.y < 0. However, the y coordinate value of VP does not verify the stair region from all the scenarios. This paper investigates and justifies this problem utilizing the exper imental analysis and introduces a mathematical model to estimate the location of the VP of the stair region. Finally, support vector machine (SVM) classifier is utilized instead of VP to recognize the stair can didate region and the performance of SVM is compared with respect to the VP. For that, rotational invari ant uniform local binary pattern (LBP) is used for feature extraction. Stair images captured under different orientation and illumination conditions have been used to test the proposed method to evaluate the resultant accuracy.