The motive of this Traffic Police Hand Gesture Recognition System for self reliant cars is constructed the use of the MoveNet Thunder version from TensorFlow`s MoveNet. It locates 3 most important gestures regularly utilized by visitors police in India, namely `Stop', 'Turn Left', and 'Move Forward'. To boom robustness, the gadget is skilled o n a custom dataset of 8000 pix taken in special environments, and neural community structure at the full-size quantity of information include Dense and dropout layers to maximise accuracy and keep away from overfitting. Furthermore, actual time face detection is used, wherein the officer is best recorded if he's looking the camera, and Haar cascades are used for that. The version is capable of acquire 89% accuracy, sturdy category overall performance throughout maximum gesture training are observed; but mild misclassifications took place among comparable gestures and became demonstrated with the CARLA simulator below everyday and climate and lights various conditions. Promising outcomes are tested through the prototype for secure and green integration into the self reliant car navigation in visitors managed environments.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Enhancing Autonomous Vehicle Navigation: Traffic Police Hand Gesture Recognition for Self-Driving Cars in India using MoveNet Thunder


    Contributors:


    Publication date :

    2025-06-17


    Size :

    592731 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English






    Enhanced BiLSTM for Traffic Police Gesture Recognition in Autonomous Vehicles

    Sunil, Vaidya Pranay / Anjali, A. / Devangan, Deepak Kumar | IEEE | 2024