Final Year Project Title: Implementation of walking aid for Visually Impaired people on Accelerated System (Jetson Nano)

Objectives we are able to achieve:

• Annotated the dataset of Images using CVAT (Computer Vision Annotation Tool)

• Achieved to train the SSD (Single Shot Detector) Computer Vision Algorithm for Facial and Object Detection.

• Achieved to do auto-labeling of the captured Images.

• Achieved made a location sender notification system in case of Emergency for the visually impaired person

Data on which I have trained this model is in the following link:

https://drive.google.com/drive/folders/1SXUKru6zTs9BHBOtAS_4GNyW9Lx4pB4i?usp=sharing

Trained Model files are in the following link and the best model file with less loss is converted into ONNX format:

https://drive.google.com/drive/folders/1ExNsCVFC9VIQEhqDTUWdpsMhbf05WNtw?usp=sharing

Image of the Mechanical Structure of the Project:

Accuracy of the Model is around 85%, but the accuracy can be improved by increasing the dataset and the number of Epochs.

Thank you for reading the ReadMe File really appreciate that!

Have a great day and Take Care.

Regards,

Saif