The problem Smart-i solves

Everyday there is a rising number of tragic road accidents, public harassment and other violent offenses. Most of these events, especially most of the accidents are caused by others irresponsiblity. One of the possible sources of evidence is the CCTV footage of the incidents.But there is a greater chance for the tampering of these evidences by influential people with money and power. This denies justice to the victim. We focus on providing equal justice to everyone by safeguarding the evidence. In the current system, CCTV footage is often stored in centralized systems, which can be vulnerable to security breaches and data loss. The possibility of tampering with crucial evidence is high since it is submitted in any kind of document.

Smart-i is a blockchain-based evidence management system to prevent manipulation of surveillance camera footage and crime investigation documents. Road accidents, harassments etc are detected from real time surveillance camera footage using a machine learning algorithm and snapshots of these videos are uploaded into a decentralized storage. Upon detection of an accident, the footage will be securely stored in a blockchain to provide tamper-proof evidence. Detection of road accidents can also act as a way to respond to emergencies more quickly. Notifications can be sent to the concerned authorities for immediate help. This technology has the potential to greatly improve public safety and provide crucial information for accident investigations. Smart-i can also be used in the future to find the missing persons, identify suspicious activities etc. By leveraging the immutability and security of blockchain technology, the project aims to provide a reliable and efficient system for accident detection and documentation.

Challenges we ran into

Accuracy in the Machine Learning Models: Model has to be trained with much more datasets to make predictions accurately and data for different categories has to be trained.
Challenge during development we faced: Difficulty in storing detected video to the decentralized storage in real time. However, we were able to solve this issue.