Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{195953,
        author = {M Parkavi and U Sripriyanka and V P Yoshik and S Vallisree},
        title = {FPGA-Based Real-Time Hand Gesture Recognition Using Depthwise Separable CNN for Contactless PowerPoint Control},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {8164-8171},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=195953},
        abstract = {This paper presents the design and hardware deployment of a Depth wise Separable Convolutional Neural Network (DS-CNN) for real-time hand gesture recognition, enabling fully contactless control of Microsoft PowerPoint presentations. Ten distinct hand gestures were identified and mapped to specific presentation operations, including slide navigation, cursor movement, and volume adjustment. After being trained in TensorFlow, the model achieved over 98% classification accuracy, demonstrating high generalization across a range of background and illumination conditions. The trained network was synthesized into Verilog using Vivado High-Level Synthesis (HLS) and implemented on a Xilinx Zynq-7000 FPGA in order to provide low-latency, energy-efficient inference. While PyAutoGUI converted identified gestures into real-time system instructions, Google Media Pipe retrieved 21 hand-joint landmarks each frame, which functioned as feature inputs to the classifier. The actual system used about 35% less power than a GPU baseline, maintained FPGA resource usage below 70%, and sustained 20–25 frames per second with per-gesture latency under 50 ms. These results demonstrate that lightweight depth wise separable convolution combined with FPGA hardware acceleration results in a useful, touchless HCI platform appropriate for educational and medical settings.},
        keywords = {CNN, Depth wise Separable Convolution, FPGA, gesture recognition, HLS, human–computer interaction, Media Pipe, PowerPoint automation, Zynq-70},
        month = {April},
        }