Classical-quantum transfer learning is a recent development in the field of quantum computing, which involves the modification of a pre-trained classical network and compounding it with a variational quantum circuit. This paper puts forward a quantum transfer learning-based approach for three different image classification tasks-classifying organic and recyclable from Trash, TB detection from chest X-ray images and detecting the presence of cracks from concrete crack images. The model used in this paper is a concatenation of pre-trained classical feature extractor with a quantum circuit as classifier. This paper compares the classification results obtained using various pre-trained networks such as VGG19, DenseNet169 and AlexNet, as feature extractors. From the obtained results, it is inferred that, DenseNet, Alexnet and VGG19 performs better in trash, TB and crack datasets, respectively. No model is the best for all the classification tasks, it is purely based on parameters such as dataset size, test-train split and learning rate.

No abstract

No abstract

No abstract