Proto-OOD: Enhancing OOD Object Detection with Prototype Feature Similarity
This addresses the issue of OOD object detection for models trained on limited categories, offering a domain-specific improvement.
The paper tackles the problem of neural networks mispredicting out-of-distribution (OOD) objects by proposing Proto-OOD, a method that uses prototype feature similarity for OOD detection, which significantly reduces the false positive rate (FPR) when tested with Pascal VOC as in-distribution and MS-COCO as OOD datasets.
Neural networks that are trained on limited category samples often mispredict out-of-distribution (OOD) objects. We observe that features of the same category are more tightly clustered in feature space, while those of different categories are more dispersed. Based on this, we propose using prototype similarity for OOD detection. Drawing on widely used prototype features in few-shot learning, we introduce a novel OOD detection network structure (Proto-OOD). Proto-OOD enhances the representativeness of category prototypes using contrastive loss and detects OOD data by evaluating the similarity between input features and category prototypes. During training, Proto-OOD generates OOD samples for training the similarity module with a negative embedding generator. When Pascal VOC are used as the in-distribution dataset and MS-COCO as the OOD dataset, Proto-OOD significantly reduces the FPR (false positive rate). Moreover, considering the limitations of existing evaluation metrics, we propose a more reasonable evaluation protocol. The code will be released.