One-way Shot Detector Approach Using YOLOv4 to Detect Objects For Robbery Detection
Robbery is a criminal act which has major consequences on not only the victims but also their family and friends. Our goal is to reduce the number of robberies and minimize the loss faced by the victims as a result of this heinous act. We are going to use the CCTV footage provided by the secu
2025-06-28 16:34:19 - Adil Khan
One-way Shot Detector Approach Using YOLOv4 to Detect Objects For Robbery Detection
Project Area of Specialization Artificial IntelligenceProject SummaryRobbery is a criminal act which has major consequences on not only the victims but also their family and friends. Our goal is to reduce the number of robberies and minimize the loss faced by the victims as a result of this heinous act.
We are going to use the CCTV footage provided by the security cameras and detect the crime. This can be done by detecting aggressiveness in the footage or by detecting the use of a weapon like knives or any kind of a firearm.
Project ObjectivesThe objective of this project is to implement and test the feasibility of One-way Shot Detector Approach to Detect Objects For Robbery Detection. This project is based on a research paper that proposes this approach, hence in FYP-I we trained our models and then further implemented in Fyp-II .To make an interface that takes a dataset and initial threshold values, return an evaluation score for all the objects and final thresholds after refinement. Based on the evaluation score, detects the objects of all different categories. From this we can detects all the weapons. Therefore in short, developing the One-way Shot Detector Approach to Detect Weapons that used in Robbery Detection by detecting all the objects either they are weapons or not.
Project Implementation MethodWe will use digital image processing and computer vision to detect robberies involving any weapon or a certain level of aggressiveness using the video feed provided
Benefits of the Project- Can be used to inform the authorities automatically.
- Help reduce the crime rate.
3. Help in reducing the damage faced by the victim,
Technical Details of Final Deliverable- For training we use Yolo V4 Object Detection Technique
- Trained the images of different wapons
- Take action for every image that is traIned
- And for testing weapons we gave video and model predict the wepons identity
- Then detect Robbery
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 70000 | |||
| computer | Equipment | 1 | 30000 | 30000 |
| cctv camera | Equipment | 1 | 40000 | 40000 |