Skip to content

Purpose of this project was verify the User Authentication and reduce the false positive predictions by improving the effectiveness of detection full-face helmets with computer vision and machine learning technologies. Also, send emergency alerts to the relevant authorities to take necessary actions before crime or robbery occurs in secure premi…

Notifications You must be signed in to change notification settings

dazimax/custom_object_detection_imageai_yolov3_keras_retinanet_neural_networks

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 

Repository files navigation

DETECTION OF FACE COVERS INCLUDING FULL FACE HELMETS VIA CCTV CAMERAS WITH COMPUTER VISION AND MACHINE LEARNING

The purpose of this project was to verify User Authentication and reduce false-positive predictions by improving the effectiveness of detection full-face helmets with computer vision and machine learning technologies. Also, send emergency alerts to the relevant authorities to take necessary actions before crime or robbery occurs in secure premises such as Banks, ATM, Jewellery shops, etc.

METHODOLOGY

ImageAI & Keras Frameworks will use to build a YoloV3 and RetinaNet custom data-set model base on Region Convolutional Neural Networks (R-CNN) by using computer vision and machine learning technologies.

Custom dataset model trained and validated with sample image datasets of full face helmet images

As the result, this application will detect and provide prediction result for input image frames of CCTV camera, video footage.

OBJECT DETECTION ALGORITHM

Algorithm 1 - The proposed object detection algorithm
Inputs: Video Path, Output Video Path, Minimum Percentage of Detection Probability
Outputs: Detected Objects, Email Notification send status, SMS Notification send status, Incident Log status

1. Import python libraries for object detection and send notifications
2. Initialize global variables with configuration values
3. Define send SMS alert functionality
4. Define send Email alert functionality
5. Define incident log maintain functionality
6. Define object detection functionality
7.   for end of video input image frames:
8.     process the full face object detection in neural network with custom dataset model
9.     if object detection == True:
10.      call the send SMS alert function
11.      call the send Email alert function
12.    end if
13.    call the log maintain function
14.    end for
15.Call the object detection functionality with created custom dataset model and video input path

DEMO VIDEO

https://www.youtube.com/watch?v=2g1QqlS_MHQ

DATASET MODEL & TEST DATA LINK

https://mega.nz/#F!GANhzITS!LZ2qDqKW43usfPAoFKmelA

About

Purpose of this project was verify the User Authentication and reduce the false positive predictions by improving the effectiveness of detection full-face helmets with computer vision and machine learning technologies. Also, send emergency alerts to the relevant authorities to take necessary actions before crime or robbery occurs in secure premi…

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages