Skip to content

Research Publications

  1. Facial Keypoints Detection Using CNN And Autoencoders
    International Journal of Future Generation Communication and Networking
    1 Jun 2020
    Find the publication here

    Facial keypoints detection is an essential task in computer vision and augmented reality domains. We put forth a novel solution for this problem which different than existing techniques. We implement a Convolutional Neural Network-based model with Autoencoders to detect facial keypoints. The deep structure of Convolutional Neural Networks (CNN) enables the extraction of high-level features and gives more accuracy while locating each key point. Convolutional networks are shaped to predict all the points simultaneously. With autoencoders, it is possible to boost the performance of the model. Autoencoders are data-specific and can work effectively if similar data was used while training.

  2. Multi-tiered Decentralized Logical Network for Secure File Sharing
    International Journal of Future Generation Communication and Networking
    1 Jun 2020
    Find the publication here

    The system aims to establish a decentralized network for protected file sharing. The dynamic network arrangement allows the peer with high data availability and raw performance to preside over and sustain all transmissions. This is done by organizing other fellow peers in a simple tree-like logical arrangement across multiple levels keeping stronger performing devices near root to serve data portions better. Reducing the peer dependency on a single server marks a departure from the archaic client-server model. The parts of the sought data can be acquired from other peers even after disconnection of the actual source. Circulated parts of data are signed specifically so authenticity and integrity are maintained by-design. Instead of encrypting all the portions, the system reduces network traffic and processing load by encrypting only the transfer ledger consisting of encryption keys and portion hashes. This minimizes the chance of eavesdropping and man-in-the-middle attacks by disposing off portions with signature mismatch.