Generative AI improves oral lesion image classification accuracy
Generative AI-Based Data Augmentation for Oral Lesion Classification: The PhotoMOCI Dataset and Benchmark
Computer Vision and Pattern RecognitionArtificial Intelligence
Summary
Detecting oral cancer early with photos can save lives, but it’s hard to train computer programs without enough example images. The authors created a big, well-organized collection of mouth photos called PhotoMOCI to help with this. They tested different ways to make more training images using AI, including fancy methods like GANs and diffusion models. To avoid confusing the computer, they introduced a filter that picks only the best AI-generated images that look real and match the right disease type. This smart filtering boosted the accuracy of oral lesion classifiers compared to older methods.
What this means in practice
- •For medical imaging teams: Improve oral lesion detection models by incorporating selective generative AI data augmentation with a filter that ensures realistic and class-representative synthetic images.
- •For dermatology ai developers: Use the synthetic image filtering approach to enhance training datasets and model accuracy for skin lesion image classification tasks.
Authors
Marco Parola, Mario G. C. A. Cimino, Sabrina Senatore
Abstract
Early detection of oral cancer via photographic imaging presents a promising avenue for large-scale oral cavity screening. However, the development of robust deep learning models is frequently hampered by the scarcity of high-quality, annotated datasets. To address this limitation, a novel and well-curated resource, the Photographic Multi-purpose Oral Cancer Imaging (PhotoMOCI) dataset, is introduced for developing models across multiple diagnostic tasks in oral oncology. Then, a comprehensive benchmark study was conducted to investigate how various data augmentation strategies influence the performance of image classifiers. Our analysis spans different generative AI frameworks, evaluating the efficacy of traditional methods against advanced generative approaches, including Generative Adversarial Networks (GANs) and Diffusion Models (DMs). Additionally, we propose the Synthetic Image Filter (SIF), a mechanism to select specific samples based on two auxiliary models: Synthetic Proxy Classifier to ensure samples are representative of the target class and Synthetic Image Detector to verify they appear realistic, thereby selecting only the high-utility images that contribute to improving downstream performance. Across the evaluated datasets and classifiers, the best SIF-filtered setup improves accuracy over traditional augmentation in all cases, with gains of +1.73% and +2.35% on PhotoMOCI and +2.38% and +2.08% on KOCD for ResNet50 and ViT, respectively. Our findings reveal that while the direct application of generative data augmentation may yield performance drops, the integration of SIF, considering (i) how synthetic data looks real and (ii) how it reflects the discriminative features of the belonging class, provides a simple yet effective mechanism to filter out synthetic samples that confuse the classifier during training.