FSOME++: Improving Few-Shot Anomaly Detection via Abnormal Sample Augmentation

Electric vehicles (EVs) rely on highly efficient manufacturing processes that rarely produce defective parts, posing a challenge for training vision-based anomaly detection (AD) models due to limited abnormal samples. To address this, we propose FSOME++, an improved few-shot anomaly detection (AD) framework that extends the FewSOME model by integrating a small number of augmented abnormal samples during training. This leads to improved representation learning, with a clearer separation between normal and anomalous samples in the latent space. FSOME++ is evaluated on both the MVTec public AD benchmark and a custom EV Metal Triangle dataset. Experimental results show consistent improvements in evaluation metrics, including F1 score, Balanced Accuracy, and AVC across multiple backbone architectures, with ResNet-34 achieving the best trade-off between performance and stability.