top of page

Face Mask detection - Computer Vision Project Help


Mask detection is a crucial technology that utilizes computer vision algorithms and machine learning models to identify individuals wearing or not wearing masks. Its primary objective is to help enforce public health measures aimed at preventing the spread of infectious diseases, such as COVID-19, by identifying individuals who are not adhering to mask-wearing guidelines in public spaces. In this blog post, we will explore the significance of mask detection, its advantages in promoting public health and safety, and delve into a machine learning project aimed at creating a real-time face mask detector.

Problem Statement

With the ongoing COVID-19 pandemic, it is essential to proactively monitor mask-wearing compliance in crowded areas, such as airports and metros. The project's objective is to develop a Real-Time Face Mask Detector using Convolutional Neural Networks (CNN) and OpenCV. By leveraging machine learning, the system aims to address monitoring challenges and enhance public safety in these high-density environments.


The project utilizes an artificial dataset consisting of 1376 face mask images, divided into two categories: images with masks (690) and images without masks (686). The goal is to train a deep learning model capable of accurately detecting whether a person is wearing a face mask in real-time.


  1. Image Preprocessing: This step involves preparing raw image data for analysis by applying techniques such as normalization, resizing, cropping, filtering, and augmentation. These processes enhance the quality and consistency of the image data, facilitating better feature extraction by the machine learning algorithm.

  2. Label Encoding: Label encoding is employed to convert categorical or text data into numerical format, enabling compatibility with machine learning algorithms. Each unique category is assigned a numerical label, ensuring the algorithm can process the data effectively.

  3. Splitting the Data into Training and Testing Sets: The dataset is divided into separate subsets to train and evaluate the machine learning model. The training set is used to optimize model parameters, while the testing set allows for performance evaluation on unseen data, ensuring the model's ability to generalize.

  4. Training the CNN Model with MobileNetV2: Convolutional Neural Networks (CNNs) are well-suited for image processing and classification tasks. In this project, a CNN model is trained using MobileNetV2 as the base model. MobileNetV2, designed for mobile and embedded devices, employs depthwise separable convolutions to reduce the network's parameters and computational cost while maintaining high accuracy.

  5. Evaluation of the Model: The model's performance is assessed using various metrics, such as accuracy, precision, recall, F1-score, and area under the curve (AUC). By evaluating the model on the testing set, its ability to make accurate predictions on new, unseen data can be determined.

The trained CNN model achieved an accuracy of 100% and a loss close to 0.1% during training. Additionally, the validation accuracy and validation loss were close to 100% and 0.1%, respectively. These promising results demonstrate the effectiveness of the mask detection system in accurately identifying individuals wearing or not wearing masks. By leveraging machine learning and computer vision, such systems can play a crucial role in enforcing public health measures, preventing the spread of infectious diseases, and improving overall public safety.

If you need implementation for the above problem or any of its variants, feel free to contact us.

bottom of page