ISSU

Other2021RIIMA × AimX
ISSU
  • This project was developed as part of the AimX Challenge, where it was awarded third prize.
  • ISSU is an Intelligent System for Scene Understanding designed to provide visually impaired with an impression of their surroundings.
  • Developed an embedded real-time obstacle and object detection system for assistive navigation.
  • Integrated on Jetson Nano, enabling low-power, on-device inference.
  • Acquires RGB images to detect and classify objects into 9 predefined categories.
  • Estimate distance to each detected object to evaluate obstacle proximity
  • Maps the Classified objects with their distance onto a 5×5 spatial grid, providing structured scene awareness.
    • Tech Stack

    Python,OpenCV,Scikit-learn,TensorFlow,Keras,Pytorch.

    Methods

    Features Extraction,Semantic Labeling,Scene Synthesis,Deep Learning,Assistive Systems,Object Detection,Obstacle Avoidance,NVIDIA Jetson,YOLO.

    Live Preview
    View Code