16 October 2017 in Journal of Biophotonics
27 March 2023 in Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies
In clinical settings, most automatic recognition systems use visual or sensory data to recognize activities. These systems cannot recognize activities that rely on verbal assessment, lack visual cues, or do not use medical devices. We examined speech-based activity and activity-stage recognition in a clinical domain, making the following contributions. (1) We collected a high-quality dataset representing common activities and activity stages during actual trauma resuscitation events-the initial evaluation and treatment of critically injured patients. (2) We introduced a novel multimodal network based on audio signal and a set of keywords that does not require a high-performing automatic speech recognition (ASR) engine. (3) We designed novel contextual modules to capture dynamic dependencies in team conversations about activities and stages during a complex workflow. (4) We introduced a data augmentation method, which simulates team communication by combining selected utterances and their audio clips, and showed that this method contributed to performance improvement in our data-limited scenario. In offline experiments, our proposed context-aware multimodal model achieved F1-scores of 73.2±0.8% and 78.1±1.1% for activity and activity-stage recognition, respectively. In online experiments, the performance declined about 10% for both recognition types when using utterance-level segmentation of the ASR output. The performance declined about 15% when we omitted the utterance-level segmentation. Our experiments showed the feasibility of speech-based activity and activity-stage recognition during dynamic clinical events.
21 December 2024 in Computer Vision and Image Understanding
WGWaverly V. Gestrich-Thompson
Waverly V. Gestrich-Thompson
+1
20 November 2023 in arXiv
In supervised speech separation, permutation invariant training (PIT) is widely used to handle label ambiguity by selecting the best permutation to update the model. Despite its success, previous studies showed that PIT is plagued by excessive label assignment switching in adjacent epochs, impeding the model to learn better label assignments. To address this issue, we propose a novel training strategy, dynamic sample dropout (DSD), which considers previous best label assignments and evaluation metrics to exclude the samples that may negatively impact the learned label assignments during training. Additionally, we include layer-wise optimization (LO) to improve the performance by solving layer-decoupling. Our experiments showed that combining DSD and LO outperforms the baseline and solves excessive label assignment switching and layer-decoupling issues. The proposed DSD and LO approach is easy to implement, requires no extra training sets or steps, and shows generality to various speech separation tasks.
16 November 2026 in ACM Transactions on Computing for Healthcare
In this study, we present a proof-of-concept system exploring the feasibility of automatically capturing physiological values during the primary survey phase of trauma resuscitation—a complex, time-sensitive evaluation of severely injured patients. We developed and evaluated a speech-driven activity recognition system designed to detect and extract key physiological parameters (e.g., age, manual blood pressure, Glasgow Coma Scale) from spoken clinical dialogue in real-time. By using audio recordings from actual trauma resuscitation cases, our system performs streaming inference to identify verbalized measurements associated with primary survey activities. Experimental results show that our approach can recognize seven attributes with an average accuracy of 72% in a challenging clinical environment with acceptable latency. We present this work as an initial feasibility demonstration rather than a deployment-ready system, highlighting both the opportunities and the remaining challenges of automating documentation in fast-paced emergency environments.