Generative Active Learning with Variational Autoencoder for Radiology Data Generation in Veterinary Medicine

Recently, with increasing interest in pet healthcare, the demand for computer-aided diagnosis (CAD) systems in veterinary medicine has increased. The development of veterinary CAD has stagnated due to a lack of sufficient radiology data. To overcome the challenge, we propose a generative active learning framework based on a variational autoencoder. This approach aims to alleviate the scarcity of reliable data for CAD systems in veterinary medicine. This study utilizes datasets comprising cardiomegaly radiographic image data and chronic kidney disease ultrasound image data. After removing annotations and standardizing images, we employed a framework for data augmentation, which consists of a data generation phase and a query phase for filtering the generated data. The experimental results revealed that as the data generated through this framework was added to the training data of the generative model, the frechet inception distance decreased from 84.14 to 50.75 in the radiographic image and from 127.98 to 35.16 in an ultrasound image. Subsequently, when the generated data were incorporated into the training of the classification model, the true negative of the confusion matrix also improved from 0.16 to 0.66 on the radiograph and from 0.44 to 0.64 on the ultrasound image. The proposed framework has the potential to address the challenges of data scarcity in medical CAD, contributing to its advancement.