Git Product home page Git Product logo

crime-prediction-deep-learning's Introduction

Project Interstice

Introduction

Surveillance is an integral part of security and control. The job requires extended periods of looking into multiple video footages for an undesirable incident or evidence following the incident, which proves to be highly inefficient considering the massive amount of footages. This manual task provides evidence in court but is rarely used to prevent crime or react to it in real-time.

🔭   About this Project

In this project, we aim to develop a real-time crime detection technology that can be integrated with any security system, to ensure public safety through visual crowd surveillance.
The system will detect and simultaneously alert any violent activity captured by the camera. The technologies encompasses domains of Computer Vision and Deep Learning models like CNN, R-CNN, LSTM. The dataset consists of 500+ videos scraped off internet and categorized into violent and non-violent activities.

🛠  Tech Stack

Tensorflow  Django  Python  HTML  CSS 

💼   Working

Though our major aim is to implement this system for Smart Monitoring, we have also developed a micro-website based on Django to show the usefulness and effectiveness of this project.
To use the website, please follow the following steps: cd Website

  • Open the directory labelled website using command line

    cd Visionary/Website

  • Run the Django server by:

    python manage.py runserver


Note : This Website required OpenCV, Numpy, Django and Tensorflow installed on the host machine. To install these dependencies, run the following command

pip3 install -r requirements.txt


✍️   Current Progress

We tried and tested various architectures for this project. A brief summary and stats for each is given below:

  1. Customised CNN Model: The customised CNN model with optimised parameters performed well on the training and sufficiently good in the test dataset. The architecture is currently being used by the website and test.ipynb file.

    • Accuracy on training: 0.7990
    • Loss on training: 0.4292
    • Val_loss: 0.7605
    • Val_accuracy: 0.6680
    • Optimizer: Adam
    • Loss: BinaryCrossEntropy
    • Epochs: 50
    • steps_per_epoch: 50



  2. VGG Net: VGGNet Architecture displayed an accuracy of 60% on training and 55% on testing dataset.

    • Testing Accuracy: 0.5513



  3. AlexNet: AlexNet showed accuracy of 57% on training and a similar accuracy on the testing dataset.

    • Testing Accuracy: 0.5729



  4. Inception+CustomisedCNN: Using transfer learning of Inception Architecture and passing it to CustomisedCNN trained the model with satisfactory results. Link to model file : Google Drive Link - Training_accuracy: 89% - Validation_accuracy: 76% - Epochs: 30 - Steps per Epoch: 100 - Optimizer: RMSprop with LR 0.0001 - Loss: Binarycrossentropy



⚙️   Ultimate Objective

The model uses the technique of Multiple Object Detection with Localization tracking the movement of people and then categorizing it into violent or non-violent behavior. Our proposed solution will ensure Public Safety and security without any human toil with an instant alert to the concerned authorities. Through Crowd Monitoring and Behavioral Analysis our solution aims to evoke the sense of security in men and women. The constant rise of criminal activities, their unexpectedness, and scope of harm that can be inflicted can be exponentially reduced through our proposed system.

Link to current protoytpe : Click Here py -3 manage.py runserver "# Crime-Prediction-Deep-Learning" "# Crime-Prediction-Deep-Learning" "# Crime-Prediction-Deep-Learning" "# Crime-Prediction-Deep-Learning" "# Crime-Prediction-Deep-Learning"

crime-prediction-deep-learning's People

Contributors

sanjanag05 avatar rajkaranm avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.