Fighting the Currency Theft Problem in Nigeria Churches through Computer Vision

Local Chapter Abakaliki, Nigeria

Coordinated byNigeria ,

Status: Completed

Project Duration: 17 Dec 2022 - 12 Jan 2023

Open Source resources available from this project

Project background.

Studies have shown that about 30% of the donations made in churches during church services through the regular tithe and offering are remitted to the church authority by those church workers in charge of it. This act of treasury theft has made most churches unable to run their activities of churches effectively. Efforts made by some church heads to combat this ugly menace have yielded little or no result, hence the need for a computer vision approach to solving the problem.

The problem.

In churches there, tithes and offerings are dropped inside an offering box by the congratulations. These boxes are enclosed boxes just like an election ballot box. After the church service, some delegated church workers collect the box, open it, sort the money therein, count it and forward the alleged total sum to the head of the church who will probably take the money to the bank. So during that process of sorting and counting, some part of the donated money is not accounted for. For example, the total sum may be 10,000 nairas but they will only remit 6,000 and pocket 4,000. So the aim of this project is to develop a computer vision device that will sort, count, and display the total sum in the box in real time.

Project goals.

The creation of automated currency recognition and counting system. In the end, a device can detect the currency note as it is being dropped into the offering box, classify them into their denominations, and calculate and display the total sum in a dashboard in real-time.  - Source for Naira Notes of all denominations( about 1000 images each) - Carryout data preprocessing. - Develop a model using ResNet50 or any Deep learning algorithms. - Carry out inference with the trained model using test data. - Develop an embedded system (computer vision) that will serve as an offering box( Note: this will probably be in the next challenge as a continuation).

Project plan.

  • Week 1

    Data Collection (pre-week 1 even)
    Data Pre-Processing

  • Week 2

    Data Pre-Processing

  • Week 3

    Exploratory Data Analysis Modelling

  • Week 4

    Modelling (cont)

  • Week 5

    Deep Learning Model testing with new data

Learning outcomes.

1. Collection of Data.
2. Data Cleaning.
3. Data Analysis.
4. Deep Learning of Naira Notes.

Share project on: