Design and implementation of quantum hippo inspired convolutional neural networks using parametric quantum circuits for an efficient lung cancer classification

With the advancement of Artificial Intelligence in medical and engineering fields, the most unique solutions were deployed in any individual’s life for prolonging their life span against the odds of growing disorders. One such disorder is Lung Cancer which is predominantly found in both women and men causing a disturbed life cycle which could lead to mental stress and even fatal end when unnoticed. In recent times, computer-aided diagnostic (CAD) act as a major automating diagnosing tool by building the self-tailored learning algorithms founded on Classical Machine and Deep Learning model. However, training classical learning frameworks consumes a huge computational resources which leads to complexity and low diagnostic performance. To overcome this problem, Quantum Hippo Optimized Convolution Neural Network (QHO-CNN) has been proposed to homogenize the quantum computations on classical computers which can effectively diagnose lung cancers with the high-speed computations. The proposed learning framework consists of four components namely Data collection & Data pre-processing, Classical Hippo Optimized Convolutional Neural networks, and Quantum based model using parametric quantum circuits (PQC), Evaluation and Analysis. The extensive experimentation carried out using LIDC-IDRI Lung cancer datasets which consists of original 1018 CT Lung Images and various learning capability tests was performed, which is then compared with the other learning framework. Results demonstrate that quantum-based learning framework has produced the accuracy of 0.97, precision of 0.964, recall of 0.963 and F1-score of 0.97 besides showing its strength of success in terms of recognising the image data and quantum training(5.431 HRS) against the other existing quantum models.