Ensuring the reliability of open-world intelligent systems heavily relies on effective out-of-distribution (OOD) detection. Despite notable successes in existing OOD detection methods, their performance in scenarios with limited training samples is still suboptimal. Therefore, we first construct a comprehensive few-shot OOD detection benchmark in this paper. Remarkably, our investigation reveals that Parameter-Efficient Fine-Tuning (PEFT) techniques, such as visual prompt tuning and visual adapter tuning, outperform traditional methods like fully fine-tuning and linear probing tuning in few-shot OOD detection. Considering that some valuable information from the pre-trained model, which is conducive to OOD detection, may be lost during the fine-tuning process, we reutilize features from the pre-trained models to mitigate this issue. Specifically, we first propose a training-free approach, termed uncertainty score ensemble (USE). This method integrates feature-matching scores to enhance existing OOD detection methods, significantly narrowing the gap between traditional fine-tuning and PEFT techniques. However, due to its training-free property, this method is unable to improve in-distribution accuracy. To this end, we further propose a method called Domain-Specific and General Knowledge Fusion (DSGF) to improve few-shot OOD detection performance and ID accuracy under different fine-tuning paradigms. Experiment results demonstrate that DSGF enhances few-shot OOD detection across different fine-tuning strategies, shot settings, and OOD detection methods. We believe our work can provide the research community with a novel path to leveraging large-scale visual pre-trained models for addressing FS-OOD detection. The code will be released.

Many young people have suffered from cervical spondylosis in recent years due to long-term desk work or unhealthy lifestyles. Early diagnosis is crucial for curing cervical spondylosis. The Cobb angle method is the most common method for assessing spinal curvature. However, manually measuring the Cobb angle is time-consuming and heavily dependent on personal experience. In this paper, we propose a fully automatic system for measuring cervical spinal curvature on X-rays using the Cobb angle method, which can reduce the workload of clinicians and provide a reliable basis for surgery. Hybrid transformer network (HTN) blends a self-attention mechanism, self-supervision learning, and feature fusion. In addition, a new cervical spondylosis dataset is proposed to evaluate our method. Our model can achieve a SMAPE of 11.06% and a significant Pearson correlation coefficient of 0.9619 (p < 0.001) on our dataset. The absolute difference between the ground truth and the prediction obtained is less than 2°, implying clinical value. Statistical analysis proves the reliability of our method for Cobb angle estimation. To further prove the validity of our method, the HTN was also trained and evaluated on the public AASCE MICCAI 2019 challenge dataset. The experimental results show that our method can achieve comparable performance to state-of-the-art methods, which means that our method can measure the curvature of the neck and the entire spine.

The existing research on open-set recognition (OSR) and novel class discovery (NCD) has largely treated these tasks as independent fields. OSR aims to identify samples that do not belong to the training set classes, while NCD seeks to further classify such unseen, unlabeled samples into novel classes. However, there is a lack of a unified framework to automate both tasks systematically. In this paper, we propose a unified training framework to identify and categorize unseen samples. Specifically, we conduct a comprehensive evaluation of existing post hoc OSR methods and observe that their performance is highly sensitive to the temperature scaling factor. To address this, we introduce a distance-based evaluation method for OSR, which not only outperforms existing post hoc approaches but also integrates seamlessly with them to deliver enhanced performance. Furthermore, we developed a prototype-based classification head leveraging this distance metric, which facilitates compact feature representations for known classes and guides the clustering of unknown classes, thereby significantly enhancing the classification accuracy for novel classes. On the CUB-200-2011 dataset, our unified framework achieves a 0.95–6.12% improvement in AUROC scores on OSR benchmarks and a 3.19% increase in classification accuracy for novel classes. Extensive experiments and visualizations validate the effectiveness of the proposed approach. We believe that this unified framework will pave the way for automating the integration of OSR and NCD, offering a more efficient and systematic approach to addressing these tasks.