Automatic phase prediction from low-level surgical activities

Forestier, Germain; Riffaud, Laurent; Jannin, Pierre

doi:10.1007/s11548-015-1195-0

Cited by 38 publications

(47 citation statements)

References 20 publications

(27 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similar to our work in activity prediction, we use a history h of the last n activities and relate it to the phase the most current activity h [26]. This is similar to the idea of a "local context" by Forestier et al [27]. The inclusion of a history is important because the meaning of an activity depends on surrounding activities.…”

Section: Experience-based Approach For Phase Recognition Using Randommentioning

confidence: 99%

Bridging the gap between formal and experience-based knowledge for context-aware laparoscopy

et al. 2016

View full text Add to dashboard Cite

show abstract

Section: Experience-based Approach For Phase Recognition Using Randommentioning

confidence: 99%

Bridging the gap between formal and experience-based knowledge for context-aware laparoscopy

et al. 2016

View full text Add to dashboard Cite

show abstract

“…However, these sensors could interfere with the work and impact patient safety. Manually recorded activity logs were used to predict the surgical phase by using a decision tree structure [4]. However, manual activity logging is subjective, requires expert knowledge, and is labor-intensive.…”

Section: Realted Workmentioning

confidence: 99%

“…This kind of modeling is used for medical education, evaluating team performance, operation planning, and task detection. However, to create high-performance decision support systems, it is necessary to have contextual awareness of the performed activities in relation to the entire process [2][3][4]. To detect and predict process phase, previous research has proposed methods using different data sources.…”

Section: Introductionmentioning

confidence: 99%

“…To avoid the labor-intensive manual logging of human observations, sensors (including cameras, wearable sensors, and machine recordings) have been used for data collection. Tool usage recordings [5], medical equipment signals [6], medical event logs [4][7], body-worn sensors [3], and multimodal Kinect sensors [1] have previously provided input data for surgical phase detection. However, human language has been overlooked as an input data source for process modeling.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Language-Based Process Phase Detection in the Trauma Resuscitation

Chen

et al. 2017

2017 IEEE International Conference on Healthcare Informatics (ICHI)

View full text Add to dashboard Cite

Process phase detection has been widely used in surgical process modeling (SPM) to track process progression. These studies mostly used video and embedded sensor data, but spoken language also provides rich semantic information directly related to process progression. We present a long-short term memory (LSTM) deep learning model to predict trauma resuscitation phases using verbal communication logs. We first use an LSTM to extract the sentence meaning representations, and then sequentially feed them into another LSTM to extract the meaning of a sentence group within a time window. This information is ultimately used for phase prediction. We used 24 manually-transcribed trauma resuscitation cases to train, and the remaining 6 cases to test our model. We achieved 79.12% accuracy, and showed performance advantages over existing visual-audio systems for critical phases of the process. In addition to language information, we evaluated a multimodal phase prediction structure that also uses audio input. We finally identified the challenges of substituting manual transcription with automatic speech recognition in trauma resuscitation.

show abstract

“…Different networks were trained for process-phase detection and for activity recognition. The experimental results showed that our system achieved performance on phase detection comparable to the system that used manually-generated log of executed tasks as input to phase recognition [10]. Our system recognized 10 common medical activities directly from RFID data with F -score 18% greater than an existing RFID-based system in the same application scenario [6].…”

Section: Introductionmentioning

confidence: 99%

Deep Learning for RFID-Based Activity Recognition

Zhang

Marsic

et al. 2016

Proceedings of the 14th ACM Conference on Embedded Network Sensor Systems CD-ROM

126

View full text Add to dashboard Cite

We present a system for activity recognition from passive RFID data using a deep convolutional neural network. We directly feed the RFID data into a deep convolutional neural network for activity recognition instead of selecting features and using a cascade structure that first detects object use from RFID data followed by predicting the activity. Because our system treats activity recognition as a multi-class classification problem, it is scalable for applications with large number of activity classes. We tested our system using RFID data collected in a trauma room, including 14 hours of RFID data from 16 actual trauma resuscitations. Our system outperformed existing systems developed for activity recognition and achieved similar performance with process-phase detection as systems that require wearable sensors or manually-generated input. We also analyzed the strengths and limitations of our current deep learning architecture for activity recognition from RFID data.

show abstract

Automatic phase prediction from low-level surgical activities

Abstract: We show that using the local context allows us to improve the results compared with methods only considering single activity. Experiments show that the use of the local context makes our method very robust to noise and that clustering the input data first improves the predictions.

Cited by 38 publications

References 20 publications

Bridging the gap between formal and experience-based knowledge for context-aware laparoscopy

Bridging the gap between formal and experience-based knowledge for context-aware laparoscopy

Language-Based Process Phase Detection in the Trauma Resuscitation

Deep Learning for RFID-Based Activity Recognition

Contact Info

Product

Resources

About