Abstract: Quantifying deep neural classifier confidence is difficult but essential for safety-critical applications. This paper introduces the true class probability (TCP) as a model confidence target criterion. TCP outperforms maximum class probability (MCP) for confidence…
Abstract: Autonomous driving requires accurate and fast point cloud 3D object detection. One-stage 3D object detection methods can achieve real-time performance, but they are dominated by anchor-based detectors that are inefficient and require post-processing. This…
Abstract: Biomedical Relation Extraction (RE) systems classify biomedical entities to improve our understanding of biological and medical processes. Most cutting-edge systems use deep learning to find relationships between similar entities like proteins or drugs. Ontologies,…
Abstract: Highway and main road cracks are common. Manual road crack evaluation is tedious, inaccurate, and difficult to implement. The complex ambient conditions—illumination, shadow, dust, and crack shape—make the computer vision-based solution difficult. Most cracks…
Abstract: Background: Pacing leads with miniaturized accelerometers monitor myocardial function. This requires acceleration signal functional indices. A method that automatically detects aortic valve opening (AVO) and closure (AVC) will aid such extraction. We used high-fidelity…
Abstract: The most common thoracic cage abnormality is pectus excavatum (PE), which is assessed using three CT indices: Haller, correction, and asymmetry. Manually performing this analysis is time-consuming and variable. This paper proposes a fully…
Abstract: AutoAtlas, a novel neural network architecture, partitions and learns 3D brain MRI volumes unsupervised. Two neural networks perform multi-label partitioning based on local texture in the volume and compress the information within each partition…
Abstract: BAA can diagnose endocrine and metabolic disorders in children. Deep learning-based bone age classification methods use global images or local information from bounding boxes or key points. Training with the global image underutilizes discriminative…
Abstract: Deep neural networks increasingly use attention to interpret and improve model performance, but little research has examined how attention progresses to complete a task and whether it is reasonable. We propose an Attention with…
Abstract: This paper generalizes the Attention in Attention (AiA) mechanism in P. Fang et al., 2019 by using explicit mapping in reproducing kernel Hilbert spaces to generate input feature map attention values. Inner and outer…
Abstract: Clinical criteria—bradykinesia, rest tremor, rigidity—diagnose Parkinson’s disease (PD). Clinical rating scales for PD severity are subject to inter-rater variability. This paper proposes a deep learning-based automatic PD diagnosis method using videos for clinical use….
Abstract: Colonoscopy has developed many automated image analysis methods in recent decades. Several recent multi-center clinical trials have tested real-time implementation of the most promising colonoscopy methods. All trials suggest colorectal cancer prevention. We review…
WhatsApp us