A study on computer vision for facial emotion recognition.

Artificial intelligence has been successfully applied in various fields, one of which is computer vision. In this study, a deep neural network (DNN) was adopted for Facial emotion recognition (FER). One of the objectives in this study is to identify the critical facial features on which the DNN model focuses for FER. In particular, we utilized a convolutional neural network (CNN), the combination of squeeze-and-excitation network and the residual neural network, for the task of FER. We utilized AffectNet and the Real-World Affective Faces Database (RAF-DB) as the facial expression databases that provide learning samples for the CNN. The feature maps were extracted from the residual blocks for further analysis. Our analysis shows that the features around the nose and mouth are critical facial landmarks for the neural networks. Cross-database validations were conducted between the databases. The network model trained on AffectNet achieved 77.37% accuracy when validated on the RAF-DB, while the network model pretrained on AffectNet and then transfer learned on the RAF-DB results in validation accuracy of 83.37%. The outcomes of this study would improve the understanding of neural networks and assist with improving computer vision accuracy.

Impact of preoperative carbohydrate loading on gastric volume in patients with type 2 diabetes.

BACKGROUND: Enhanced recovery after surgery advocates that consuming carbohydrates two hours before anesthesia is beneficial to the patient's recovery. Patients with diabetes are prone to delayed gastric emptying. Different guidelines for preoperative carbohydrate consumption in patients with diabetes remain controversial due to concerns about the risk of regurgitation, aspiration and hyperglycemia. Ultrasonic gastric volume (GV) assessment and blood glucose monitoring can comprehensively evaluate the safety and feasibility of preoperative carbohydrate intake in type 2 diabetes (T2D) patients. AIM: To evaluate the impact of preoperative carbohydrate loading on GV before anesthesia induction in T2D patients. METHODS: Patients with T2D receiving surgery under general anesthesia from December 2019 to December 2020 were included. A total of 78 patients were randomly allocated to 4 groups receiving 0, 100, 200, or 300 mL of carbohydrate loading 2 h before anesthesia induction. Gastric volume per unit weight (GV/W), Perlas grade, changes in blood glucose level, and risk of reflux and aspiration were evaluated before anesthesia induction. RESULTS: No significant difference was found in GV/W among the groups before anesthesia induction (P > 0.05). The number of patients with Perlas grade II and GV/W > 1.5 mL/kg did not differ among the groups (P > 0.05). Blood glucose level increased by > 2 mmol/L in patients receiving 300 mL carbohydrate drink, which was significantly higher than that in groups 1 and 2 (P < 0.05). CONCLUSION: Preoperative carbohydrate loading < 300 mL 2 h before induction of anesthesia in patients with T2D did not affect GV or increase the risk of reflux and aspiration. Blood glucose levels did not change significantly with preoperative carbohydrate loading of < 200 mL. However, 300 mL carbohydrate loading may increase blood glucose levels in patients with T2D before induction of anesthesia.

[Study on the practice and effect of target teaching method in teaching of dental technology interns].

PURPOSE: To standardize the teaching process of interns in dental technology so as to make the learning progress and process goals more clearly and improve the teaching quality. METHODS: Thirty-two junior interns were selected from medical colleges and universities from 2014 to 2017, based on same learning ability, learning attitude, learning achievement and hands-on ability ，they were randomly divided into 2 groups. The experimental group adopted the goal teaching method, using relevant teaching materials, applying theory to practice closely, and trying to standardize practice. The control group entered the production lines directly without teaching materials, the students were all owed only to see, think and manipulate. The data were analyzed with SPSS 18.0 software package. RESULTS: The exam scores of the students in the experimental group were significantly higher than that of the students in the control group. Moreover, the satisfaction with the teaching methods of the experimental group was significantly higher than that of the control group (P<0.05). CONCLUSIONS: The target teaching method is effective in teaching dental technology. Students clearly understand the study progress, process goals and their operational performance is significantly improved.

Comparison of Continuous Noninvasive Blood Pressure Monitoring by TL-300 With Standard Invasive Blood Pressure Measurement in Patients Undergoing Elective Neurosurgery.

BACKGROUND: Intraoperative blood pressure (BP) is one of the basic vital signs monitoring. Compared with standard invasive BP measurement, TL-300 allows for a continuous and beat-to-beat noninvasive intraoperative BP monitoring. The current retrospective study compared the accuracy and precision of this noninvasive technique for continuous BP monitoring with that of standard invasive BP measurement in patients undergoing elective neurosurgery. MATERIALS AND METHODS: BP records of 23 patients undergoing elective neurosurgery, measured by both noninvasive TL-300 and invasive radial arterial catheter method, were retrospectively analyzed. Variability in BP data was analyzed by using linear regressions and Bland-Altman analysis. RESULTS: Four thousand three hundred eighty-one pairs of BP measurements from a total of 23 patients were included. The coefficient of determination of systolic, diastolic, and mean BP were 0.908, 0.803, and 0.922, respectively. And their bias was found to be 1.3±5.87 mm Hg (95% limits of agreement: -10.2 to +12.8 mm Hg), 2.8±6.40 mm Hg (95% limits of agreement: -9.8 to +15.3 mm Hg), and 1.8±4.20 mm Hg (95% limits of agreement: -6.4 to +10.1 mm Hg), respectively. CONCLUSIONS: TL-300 system is a promising noninvasive alternative to the invasive arterial catheter method for intraoperative BP monitoring, with a high accuracy and precision. With the limitation of the current retrospective study, further prospective method comparison studies are needed.

Neuromuscular effect of dexmedetomidine on sevoflurane: an open-label, dose-escalation clinical trial.

BACKGROUND: Sevoflurane presents reliable central neuromuscular effects. However, little knowledge is available regarding the interaction between sevoflurane and demedetomidine. We evaluated the neuromuscular effect of dexmedetomidine on sevoflurane in patients with normal neuromuscular transmission and calculated the 50% effective concentration (EC50). METHODS: One-hundred and forty-four ASA grade I~II patients with normal neuromuscular transmission, aged 20~60 years old, undergoing lower limbs surgery were enrolled in this open-label, dose-escalation clinical trial. Patients were randomly assigned into 12 groups. Each patient received intravenous 0, 0.5, or 1.0 µg/kg dexmedetomidine 15 min after inhaling 0.7, 1.0, 1.4, or 2.0 MAC sevoflurane. Neuromuscular monitoring was recorded from the adductor pollicis muscle by using acceleromyography with train-of-four (TOF) stimulation of the ulnar nerve (2 Hz every 20 s). TOF ratio was recorded before inhaling sevoflurane, 15 min after keeping constant at target MAC of sevoflurane, 30 min after receiving target dose of dexmedetomidine, and 15 min after sevoflurane washing out. RESULTS: Sevoflurane produced a concentration-dependent decrease in TOF ratio. Mean TOF ratio in 0.7, 1.0, 1.4, and 2.0 MAC groups was 97.9%, 94.9%, 84.7%, and 77.2%, respectively. Neuromuscular EC50 of sevoflurane was 1.31 MAC (95% CI: 1.236~1.388 MAC). Intravenous 0.5 and 1.0 µg/kg dexmedetomidine decreased 3.1% (EC50: 1.27 MAC [95% CI: 1.206~1.327 MAC]) and 10.7% (EC50: 1.17 MAC [95% CI: 1.122~1.217 MAC]) of neuromuscular EC50, respectively. CONCLUSIONS: Sevoflurane has a concentration-dependent central neuromuscular effect in patients with normal neuromuscular transmission. Intravenous dexmedetomidine dose-dependently decreases the neuromuscular EC50 of sevoflurane.