PESA R-CNN: Perihematomal Edema Guided Scale Adaptive R-CNN for Hemorrhage Segmentation.

Intracranial hemorrhage (ICH) is a type of stroke with a high mortality rate and failing to localize even minor ICH can put a patient's life at risk. However, its patterns are diverse in shapes and sizes and, sometimes, even hard to recognize its existence. Therefore, it is challenging to accurately detect and localize diverse ICH patterns. In this article, we propose a novel Perihematomal Edema Guided Scale Adaptive R-CNN (PESA R-CNN) for accurate segmentation of various size hemorrhages with the goal of minimizing missed hemorrhage regions. In our approach, we design a Center Surround Difference U-Net (CSD U-Net) to incorporate Perihematomal Edema (PHE) for more accurate Region of Interest (RoI) generation. We trained CSD U-Net to predict PHE and hemorrhage regions as targets in a weakly supervised manner and utilized its prediction results to generate RoI. By including more informative features of PHE around hemorrhage, this enhanced RoI generation allows a model to reduce the false-negative rate. Furthermore, these expanded RoIs are aligned with the Scale Adaptive RoI Align (SARA) module based on their size to prevent the loss of fine-scale information and small hemorrhage patterns. Each scale adaptively aligned RoI is processed with the corresponding separate segmentation network of Multi-Scale Segmentation Network (MSSN), which integrates the results from each scale's segmentation network. In experiments, our model shows significant improvement on dice coefficient (0.697) and Hausdorff distance (12.918), compared to all other segmentation models. It also minimizes the number of missing small hemorrhage regions and enhances overall segmentation performance on diverse ICH patterns.

Development of a human-friendly visual inspection method for painted vehicle bodies.

Visual inspections performed in the final stage of the vehicle manufacturing process are crucial for assuring the quality of painted vehicle bodies. However, lengthy visual inspections can cause fatigue and discomfort of the eyes, which may adversely affect inspection accuracy and efficiency. This study developed a new human-friendly visual inspection method for the detection of defects (e.g., scar and dent) on vehicle bodies, and compared its performance to a conventional inspection method in terms of critical fusion frequency (Hz, indirect measure of eye fatigue), defect detectability (%), and subjective satisfaction score (7-point Likert scale). The new method was devised to project bright-dark linear stripes onto the surface of the vehicle body and created emergent features (distorted stripes) where a defect existed. The critical fusion frequency of the new method decreased slightly (3.7%) after a 30-minute visual inspection task, whereas that of the conventional method dropped substantially (11.0%), which implied more severe eye fatigue. Additionally, the new method had significantly higher defect detectability (92.1%) and satisfaction score (5.8 points) than those (73.4% and 3.5 points) of the conventional method.

A Study on sEMG-Based Motor Variability and Functional Connectivity of the Upper Limb Depending on Weight Distributions in a Handle of a Cordless Stick-Type Vacuum Cleaner.

This study investigated the muscle activities, motor variability, and functional connectivity of the upper limb as a function of weight distributions in a handle of a cordless stick-type vacuum cleaner. Eighteen female college students with experience of vacuum cleaner-use participated in testing. Five handles with different centers of mass (CM) were prepared (centroid, top-rear, top-front, bottom-front, and bottom-rear), and electromyography for the muscles of the upper limb were measured during vacuuming. The results showed that the %MVC values of the Extensor Carpi Ulnaris (p = 0.0038) and Deltoid Middle (p = 0.0094) increased but that of the Biceps Brachii (p = 0.0001) decreased, as the CM moved from the top to bottom area of the handle. The motor variability of the Extensor Carpi Ulnaris (p = 0.0335) and Brachioradialis (p = 0.0394) significantly varied depending on the CM locations but failed to show significance in the post-hoc analyses. Lastly, the functional connectivity values of the muscle pairs such as the Extensor Carpi Ulnaris-Deltoid Middle (p = 0.0016), Extensor Carpi Ulnaris-Upper Trapezius (p = 0.0174), Brachioradialis-Biceps Brachii (p = 0.0356), and Biceps Brachii-Upper Trapezius (p = 0.0102) were significantly altered as a function of the CM locations. The lowest functional connectivity was found with the handle of which CM was at centroid.

Repetitive patterns in the locations of touch errors for two-thumb text entry on a smartphone.

This study investigated repetitive patterns in the locations of touch errors as a function of the shapes and positions of soft buttons on a smartphone for two-thumb text entry. Forty-three right-handed college students with smartphone-use experience were recruited for testing. An experimental application was developed, and the locations and frequencies of touch errors were measured for the button combinations of seven shapes and eight positions. More than 70.0 % of touch errors occurred within 2 mm from the boundaries of the buttons. In terms of direction, touch errors were primarily observed below the buttons, across all the button shapes and positions. Simultaneously, touch errors often appeared on the lateral sides of the buttons: (1) close to the proximal phalange of the thumbs when the buttons were placed near the initial positions of the thumbs and (2) close to the initial positions of the thumbs when the buttons were placed near the top and bottom ends of the keyboard.

Preventing Emergency Vehicle Crashes: Status and Challenges of Human Factors Issues.

OBJECTIVE: This study reports current status of knowledge and challenges associated with the emergency vehicle (police car, fire truck, and ambulance) crashes, with respect to the major contributing risk factors. BACKGROUND: Emergency vehicle crashes are a serious nationwide problem, causing injury and death to emergency responders and citizens. Understanding the underlying causes of these crashes is critical for establishing effective strategies for reducing the occurrence of similar incidents. METHOD: We reviewed the broader literature associated with the contributing factors for emergency vehicle crashes: peer-reviewed journal papers; and reports, policies, and manuals published by government agencies, universities, and research institutes. RESULTS: Major risk factors for emergency vehicle crashes identified in this study were organized into four categories: driver, task, vehicle, and environmental factors. Also, current countermeasures and interventions to mitigate the hazards of emergency vehicle crashes were discussed, and new ideas for future studies were suggested. CONCLUSION: Risk factors, control measures, and knowledge gaps relevant to emergency vehicle crashes were presented. Six research concepts are offered for the human factors community to address. Among the topics are emergency vehicle driver risky behavior carryover between emergency response and return from a call, distraction in emergency vehicle driving, in-vehicle driver assistance technologies, vehicle red light running, and pedestrian crash control. APPLICATION: This information is helpful for emergency vehicle drivers, safety practitioners, public safety agencies, and research communities to mitigate crash risks. It also offers ideas for researchers to advance technologies and strategies to further emergency vehicle safety on the road.

Glasses-type wearable computer displays: usability considerations examined with a 3D glasses case study.

This study presents usability considerations and solutions for the design of glasses-type wearable computer displays and examines their effectiveness in a case study. Design countermeasures were investigated by a four-step design process: (1) preliminary design analysis; (2) design idea generation; (3) final design selection; and (4) virtual fitting trial. Three design interventions were devised from the design process: (1) weight balance to reduce pressure concentrated on the nose, (2) compliant temples to accommodate diverse head sizes and (3) a hanger mechanism to help spectacle users hang their wearable display on their eye glasses. To investigate their effectiveness, in the case study, the novel 3D glasses adopting the three interventions were compared with two existing 3D glasses in terms of neck muscle fatigue and subjective discomfort rating. While neck muscle fatigue was not significantly different among the three glasses (p = 0.467), the novel glasses had significantly smaller discomfort ratings (p = 0.009). Relevance to Industry: A four-step design process identified usability considerations and solutions for the design of glasses-type wearable computer displays. A novel 3D glasses was proposed through the process and its effectiveness was validated. The results identify design considerations and opportunities relevant to the emerging wearable display industry.

Investigation of index finger triggering force using a cadaver experiment: Effects of trigger grip span, contact location, and internal tendon force.

A cadaver study was conducted to investigate the effects of triggering conditions (trigger grip span, contact location, and internal tendon force) on index finger triggering force and the force efficiency of involved tendons. Eight right human cadaveric hands were employed, and a motion simulator was built to secure and control the specimens. Index finger triggering forces were investigated as a function of different internal tendon forces (flexor digitorum profundus + flexor digitorum superficialis = 40, 70, and 100 N), trigger grip spans (40, 50, and 60 mm), and contact locations between the index finger and a trigger. Triggering forces significantly increased when internal tendon forces increased from 40 to 100 N. Also, trigger grip spans and contact locations had significant effects on triggering forces; maximum triggering forces were found at a 50 mm span and the most proximal contact location. The results revealed that only 10-30% of internal tendon forces were converted to their external triggering forces.

Development of a press and drag method for hyperlink selection on smartphones.

The present study developed a novel touch method for hyperlink selection on smartphones consisting of two sequential finger interactions: press and drag motions. The novel method requires a user to press a target hyperlink, and if a touch error occurs he/she can immediately correct the touch error by dragging the finger without releasing it in the middle. The method was compared with two existing methods in terms of completion time, error rate, and subjective rating. Forty college students participated in the experiments with different hyperlink sizes (4-pt, 6-pt, 8-pt, and 10-pt) on a touch-screen device. When hyperlink size was small (4-pt and 6-pt), the novel method (time: 826 msec; error: 0.6%) demonstrated better completion time and error rate than the current method (time: 1194 msec; error: 22%). In addition, the novel method (1.15, slightly satisfied, in 7-pt bipolar scale) had significantly higher satisfaction scores than the two existing methods (0.06, neutral).

Development of a usability evaluation method using natural product-use motion.

The present study developed and tested a new usability evaluation method which considers natural product-use motions. The proposed method measures both natural product-use motions (NMs) and actual product-use motions (AMs) for a product using an optical motion capture system and examines the usability of the product based on motion similarity (MS; %) between NMs and AMs. The proposed method was applied to a usability test of four vacuum cleaners (A, B, C, and D) with 15 participants and their MSs were compared with EMG measurements and subjective discomfort ratings. Cleaners A (44.6%) and C (44.2%) showed higher MSs than cleaners B (42.9%) and D (41.7%); the MSs mostly corresponded to the EMG measurements, which could indicate that AMs deviated from NMs may increase muscular efforts. However, the MSs were slightly different from the corresponding discomfort ratings. The proposed method demonstrated its usefulness in usability testing, but further research is needed with various products to generalize its effectiveness.

Effects of button position on a soft keyboard: Muscle activity, touch time, and discomfort in two-thumb text entry.

Intensive use of the thumbs for text entry on smartphones may contribute to discomfort, pain, or musculoskeletal disorders. This study investigated the effect of twenty-five button positions (5 rows × 5 columns) on a soft keyboard for two-thumb entry. Two experiments measured muscle activity, touch time, and discomfort as a function of the button positions. In Phase I, the muscle activities of two intrinsic (abductor pollicis brevis and first dorsal interossei) and two extrinsic (abductor pollicis longus and extensor digitorum communis) muscles associated with thumb motions were observed for ten college students (age: 24.2). In Phase II, touch time and discomfort were measured for 40 college students (age: 23.6). The results demonstrated that the %MVCs of the intrinsic muscles significantly increased when the thumbs flexed and abducted. Also, the button positions near the rest positions of the thumbs resulted in significantly shorter touch times (0.66 s) and lower discomfort ratings (0.70 pt) than their peripheral buttons (0.76 s; 2.29 pt).

Investigation of methods for estimating hand bone dimensions using X-ray hand anthropometric data.

This study examined two conversion methods, M1 and M2, to predict finger/phalange bone lengths based on finger/phalange surface lengths. Forty-one Korean college students (25 males and 16 females) were recruited and their finger/phalange surface lengths, bone lengths and grip strengths were measured using a vernier caliper, an X-ray generator and a double-handle force measurement system, respectively. M1 and M2 were defined as formulas able to estimate finger/phalange bone lengths based on one dimension (i.e., surface hand length) and four finger dimensions (surface finger lengths), respectively. As a result of conversion, the estimation errors by M1 presented mean 1.22 mm, which was smaller than those (1.29 mm) by M2. The bone lengths estimated by M1 (mean r = 0.81) presented higher correlations with the measured bone lengths than those estimated by M2 (0.79). Thus, the M1 method was recommended in the present study, based on conversion simplicity and accuracy.