- [Mohd Mohsin]
- [Vasu Jain]
- [Anand Chauhan]
- [Manish Raj]
Due to growing lung cancer rates and COVID-19 pandemic awareness, automated lung cancer diagnosis has garnered focus in recent years. In medical image processing, Convolutional Neural Networks (CNNs) have shown promise for accurate and efficient lung cancer identification. This research uses a lightweight CNN architecture to identify lung cancer automatically. The lightweight CNN balances model complexity with performance, making it suited for mobile devices and edge computing systems. The proposed technique is tested using a heterogeneous lung cancer dataset with chest X-ray pictures. During training, our lightweight CNN achieved 99% dataset accuracy, xceeding other similar models. The model's sensitivity and specificity—critical in lung cancer screening—reduce false negatives and positives. Precision and automated lung cancer detection may speed up early identification and care, improving patient outcomes. CNN's lightweight architecture allows real-time inference on resource-constrained devices, expanding lung cancer screening in rural and underprivileged areas. In conclusion, a lightweight convolutional neural network can identify lung cancer automatically. This approach's 99% validation accuracy on a difficult lung cancer dataset shows its potential to help doctors make accurate diagnoses. This aids lung cancer prevention and public health.
The dataset "Chest X-ray images for COVID-19" contains 3 labelled classes of 1097 X-rays images.
We will be using a custom lightweight CNN to create a classifier for the given dataset.