The tail suspension test (TST) is a widely used mouse behavioral test to evaluate the efficacy of antidepressant drugs. While the measurement of immobility, a key metric in TST, is often manually scored by a human investigator, an automated video analysis system provides more consistent and objective scoring. However, the proper validation and optimization of the automated analysis is crucial to ensure the reliability and validity of the TST results. In this study, a deep learning analysis successfully achieved accurate immobility recording using a single domestic camera observation. The system employs two deep learning models, allowing for the evaluation of temporal movement and static mouse postures. This system demonstrates immobility recognition with exceptional precision, as evidenced by a correlation coefficient (r) of 0.990 with manual annotations. The newly developed system employing deep learning models can be applied to other behavioral tests, providing an unbiased approach and contributing to advancements in various neurological research.

Changes in mouse behavior are useful for both basic and applied research. However, visual inspection by humans is subjective and time-consuming. With the advancement of deep learning, systems have been developed that are capable of automatically and quantitatively classifying mouse behavior from videos. As camera angles are typically from above or the side, consistently capturing keypoints related to limb movement can be challenging. In this study, a mouse was placed on a transparent acrylic plate and its movements were recorded from below using an RGB-D camera, successfully capturing its limbs in 3D at all times. Furthermore, by using DeepLabCut, the 3D coordinates of the mouse's keypoints were obtained. By using deep learning with the time-series data of these obtained keypoints coordinates and corresponding behavioral labels, we created a model that classifies mouse behaviors from videos. This method achieved a total accuracy of 96.7% and a walking classification accuracy of 94.5%, demonstrating higher precision compared to previous studies.

In neuroscience research, in vivo imaging techniques for mice are used to observe brain activity and link it to their behavior. Brain activity can often only be associated with observed behavioral outcomes. In other words, it is difficult to speculate on unmanifested behavior due to factors such as “hesitation” in humans. When a prediction model can predict mice behavior, if brain activity is observed in a specific brain region during incorrect predictions, that would be strong evidence of unmanifest behavior. In this study, we developed a trajectory prediction model to predict the walking trajectory of mice as a prelude to the behavior prediction model. The prediction model was applied to the behavioral analysis of mice administered an anxiolytic drug (diazepam) or saline, revealing significantly different outcomes.

Injuries during the practice of sports in gymnasiums have been reported, and one of the causes of injuries is due to environmental factors as tiny wood splinters on the gymnasium floor. Although it is important to regularly inspect gymnasium floors, it is difficult for humans to inspect the entire gymnasium floor, as it is done manually and visually, and requires a lot of time and manpower. We have developed an automatic inspection system to detect tiny splinters on the gymnasium floor. The system attaches cotton to tiny splinters and detects the attached cotton by using an image processing technique. Using this system, the entire gymnasium floor can be inspected automatically by using simply creating a 2D map. After the inspection, the system can show where splinters are located on the map. In this paper, the method for detecting splinters attached to cotton using deep learning object detection-YOLO was proposed. The detection ratio of the proposed method was improved by 25.0 % compared to the conventional method of threshold color segmentation process. In an inspection of an entire gymnasium, the proposed method detected 33 markers and was able to detect splinters that could cause injury.

Early diagnosis of diabetes peripheral neuropathy (DPN) is necessary for prevention of diabetes foot diseases in the early stage. In this study, we propose a quantitative DPN screening system which uses the change of the foot plantar images on different hardness floor based on skin hardness of patients. The purpose of this study is to estimate the severity of the DPN based on the foot plantar image. We compared the evaluation between the method of the change of the foot plantar image and classical way to calculate severity points, neurological, and hemodynamic assessments. A positive correlation (0.517) was presented between the proposed evaluation and the classical assessments. Also, we compared the area change rate with the hardness of the foot plantar skin measured by durometer and the hardness based on proposed image diagnosis. We measure the foot plantar images, the hardness, and classical assessment for 113 persons. A significant positive correlation was found between the area change rate and the hardness of the skin of big toe (0.300), first metatarsal head (0.448), and heel (0.400). A significant positive correlation (0.468) was also found between severity and skin hardness in the heel.

The paper proposes a motion planning method using a deep reinforcement learning algorithm, Asynchronous Advantage Actor-Critic (A3C). For mobile robot navigation tasks in crowds, existing path planning based approaches are limited because the surrounding environments change dynamically. The correct motion in such a dynamic environment is underspecified, and a reinforcement learning approach is suitable for generating applicable motion. We propose an A3C based motion planning method for acquiring robot motion for a robot moving through crowds. The proposed method is evaluated in simulated crowds of pedestrians. The experiment section shows the basic performance depending on training parameters and some generated motion examples in the simulator. The learning results using real pedestrian motion are also shown.

The purpose of this study is quantitative evaluation of orthodontic treatment force and moment when rotating teeth around two different axes. There is few quantitative measurements of force and moment applied to teeth during treatment. Therefore, even though the dentist performs the same way of attachment of orthodontic instrument, the movement of the teeth differ based on the dentist's treatment skill. The measurement of the orthodontic treatment force and moment is quit important, few devices can measure the force and moment. We have developed the force and moment sensing device. The device measured 3-axis forces and 3-axis moments applied to pseudo teeth model with real orthodontic instrument. The moment applied to the tooth when the bracket attached to the tooth was rotated -20 to 20 degrees around two different axes was measured. Two moments should theoretically be the same value for each angle in different axes. However, the results of measurement demonstrate that moment is different for each axis. The moment applied to the teeth varies depending on the direction of tooth rotation. The experimental results suggest that the dentist needs to consider details: the small difference of the rotation and position, and wire attachment skill.

The paper proposes a system for mapping the 3D location of a sound source using data from a microphone array, each of which gives an independent estimate of the direction. LiDAR is used to generate a 3D map of the environment and to estimate the location of the sensor in six degrees of freedom (6-DoF). By combing these modules, our system determines the direction to a sound source in 3D global space for a moving robot. From evaluation of the time series of the tracked sound stream, the position of the sound source is estimated by using the Monte Carlo localization approach. When estimated from a moving robot, an approach based on triangulation does not always adequately locate a sound source, because the estimate depends on the relationship between the location of the source and that of the moving robot. The main advantage of the proposed system is that as an audible signal is perceived, it continuously estimates the location of the source. We evaluate the system by performing an experiment using a hand-held sensor.

This paper proposes a simulation-based learning support system that animates extinct animals in synchronization with a human’s location and actions. The system provides learners with a near-real body experience, and the sense that they have entered the paleontological environment. The aim of the proposed system is for this experience to improve the learning effect. Obviously, it is impossible for learners to experience the actual paleontological environment, and it is therefore difficult for them to understand the link between this environment and its actuality. This difficulty can be decreased by the proposed system, which provides learners with a real body experience and sense of immersion. At this point in time, the authors are implementing a system based on the proposed concept. As a preliminary evaluation, we conduct an experiment using the system, and interview the participants about the improvement in the learning effect. The experimental results confirm that most learners feel a tangible improvement in the learning effect due to the enjoyment of moving their body during the learning process.

The authors perform promotion activity on science and technology for primary school children with human following robot. This robot can follow a particular person as target. In our activity, the children enjoy interaction with robot by being followed as the target. Enjoying with the robot, they are interested in the science and technology. In this paper, we present framework about the human following robot. Also we demonstrate the interaction with the robot in our promotion activity.

This paper investigates that effect of neutral voltage on shaft voltages. The shaft voltages are expected to provide cause of bearing currents. The bearing currents damage bearing in motor. We expect neutral voltages to cause shaft voltages. Thus, we construct circuit of interior motor. And we predict generation of shaft voltages by simulation using the circuit. Then we measure the neutral voltage and draw a comparison between predictive value and measured value. As a result, we can confirm connections between shaft voltages and neutral voltages.

This paper presents following continuation function for human following robot losing target at corner. To realize such function, the authors propose a human trajectory model of turning the corner. Using the proposed model, the robot can predict trajectory of the target human when he/she turns the corner and disappears from the robot's view. To move on the target trajectory, the robot can rediscovery the target and continue to follow. In order to confirm effectiveness of the proposed function and the following robot using it, the authors conduct experiment. In the experiment, the robot can continue human following. Experimental result proves the effectiveness and supports our idea successfully.

This paper proposes a method to segment out a hand grasped object from human region obtained from Kinect sensor by using body dimension database. Having dataset of human body dimensions, Multiple Regression Analysis is applied to find out the best explanatory variables for forearm and upper arm length. As a result, "body height" is selected. In order to measure "body height" accurately, Kinect depth image is utilized to search with kinematical result obtained from Kinect software. After estimating wrist position, we can segment out hand grasped region. Methods and experimental results are shown.

This paper proposes robust discriminator between bicycle rider and pusher from video sequence. There are a large number of accidents happened while riding a bicycle. The discriminator between bicycle rider and pusher is required to prevent accidents. However, previous works make no mention of the discrimination between similar shape objects such as bicycle rider and pusher. To realize the discrimination, the authors pay attention to difference of their leg movements. Bicycle rider has a pedaling movement, while pusher has a walking movement. To evaluate the difference quantitatively, we utilize CHLAC (Cubic Higher-order Local Auto-Correlation). We conduct an experiment using video sequence in order to confirm effectiveness of the proposed discriminator. In the experiment, success rate of the discrimination can be achieved more than 80%. Thus, we confirm effectiveness of the proposed discriminator.

In this paper, an automatic cancer detection method that combines multiple features to support pathologists was proposed. Cancer is the most cause of death in Japan, and patients suffering with cancer are increasing every year, while the number of pathologists is almost constant. Such issues increase the burden on the pathologists and causes service degradation for the patients. The proposed method combined three image features, Higher-order Local Auto-Correlation (HLAC) feature, Wavelet feature, Delaunay feature. At first, the features were calculated from gastric lymph node images. Then we connected each feature into each vector of varying combinations of the features, and discriminated cancer and no cancer by Support Vector Machine (SVM). HLAC, Wavelet and Delaunay features are shape, frequency, and cell-position geometrical one respectively. Cancer detection rates with more than two features combination were better than only one. In the best performance, sensitivity and specificity were 94.6% and 84.9% respectively.

We are developing a digital game, "Human SUGOROKU," in which enables game players as pieces to learn vegetation succession. In order to realize this game, it is necessary to robustly measure 3D position of people and identify them. In this paper, we proposed a method for 3D position measurement and identification of people by integrating ultrasonic sensors and Kinect sensors. The evaluation results show that proposed method is more robust than methods using ultrasonic sensor and Kinect sensor separately.

This paper describes visualizing sound pressure distribution by combining microphone array processing and camera image processing. In microphone array signal processing, capturing a sound at desired position is achieved by Delay and Sum Beam Forming (DSBF). In camera image processing, taking information of 3D position is achieved by triangulation method with stereo camera. We calculate sound level with capturing sound at 3D position taken by camera image processing. In order to visualize sound pressure distribution, calculated sound levels are superimposed on camera image. Experiments were conducted to examine visualization of sound pressure distribution by using different noise and different sound pressure level.

The detecting and tracking a specific person by using the image processing with the color camera is often influenced by illumination changes. The proposed robust detecting and tracking method to illumination changes is using the HSV color space and updating the histogram. However, the quantitative evaluating is not enough in the proposed method. In this paper, the robustness to illumination changes of the proposed method is evaluated quantitatively. Moreover, the accuracy of the tracking in the proposed method is also evaluated quantitatively. The extraction rate, the object detecting rate, and the tracking error are defined to evaluate quantitatively. The tracking error is also calculated by the MeanShift method based on the color information. The quantitatively experimental results show that the proposed method is the stable and robust tracking method.

The spherical area where sound pressure is high on a partial target made by Speaker Array is called Sound Spot. In this paper, the amount of characteristic was calculated by using HLAC for an image in video sequence. This study aims to make a peculiar person hear the sound of an individual content.

In this paper, we propose the technique classifying background area automatically by calculating the color variance in the RGB space of the picture. By this technique, the background classification robust to change of the background brightness and the background color was enabled. Moreover, we show the effectiveness of this technique by performing particle detection and verifying the result of counting. We aim at development of the particle detector which performs particle counting automatically about the “dispersion staining method” which is one of the processes of asbestos analysis. We used the pictures taken by phase contrast microscopy used for a “dispersion staining method.”

This paper presents a design concept of a novel 6-axis force sensor with simultaneous plantar deformation measurement by high-speed cameras under the sensor is presented. The sensor is based on two custom beams and some strain gages in order not to be limited a view of the high-speed cameras. In walking analysis research field, the 6-axis force and plantar skin deformation are very important factors to identify the risk of fall and to reduce risk of fall-related injuries in elderly populations. However, there were no sensor can measure 6-axis force and plantar skin deformation simultaneously. In this paper, the design of the sensor is described and perpendicular axis calibration results are shown. Furthermore, the walking experimental results show that the development sensor can be used for walking analysis.

This paper describes the Sound spot generation and evaluation by a large scale panel loudspeaker array. The authors research the sound spot by loudspeaker array. The sound spot is local high sound pressure area. We develop 128 ch panel loudspeaker array and 512 ch panel loudspeaker array to generate the sound spot. The 512 ch panel loudspeaker array is using four the 128ch panel loudspeaker array systems, and is surrounded by the square form. The authors simulate the sound spot generated by the 128 ch panel loudspeaker array and the 512 ch panel loudspeaker array. From the result of the simulation, the 128 ch panel loudspeaker array can generate straight line sound spot and the 512 ch panel loudspeaker array can generate spherical sound spot. Actual measurement of sound spot is performed by using 64 ch microphones measurement system. We confirm that each panel loudspeaker array generate three dimensional sound spot from result of measurement. This means that communication of voice information to many people is possible.

This paper describes the three-dimensional simulation and measurement of the sound pressure distribution generated by 120 ch plane loudspeaker array. The authors research the local high sound pressure distribution generated by loudspeaker array. We develop the 120 ch plane loudspeaker array. The 120 plane loudspeakers are arranged at the 120 ch plane loudspeaker array. The authors simulate the three dimensional sound pressure distribution generated by the 120 ch plane loudspeaker array. From the result of the simulation, the 120 ch plane loudspeaker array can generate three dimensional sound pressure distribution. The 120 ch plane loudspeaker array is developed based on the simulation result. Actual measurement of sound pressure distribution is performed with 64 ch microphones measurement system. We confirm that the 120 ch plane loudspeaker array can generate three dimensional local high sound pressure area from result of measurement. In addition, result of simulation is compared with result of measurement. The simulation is available for the simulation of the 120 ch plane loudspeaker array from result of comparison.

This paper describes a method of measuring a plantar skin contact area and a plantar incipient partial slip in order to clarify the relation between a plantar tactile sensibility and a stabilization mechanism for slip. The information of the plantar sense stimulus during walk had been not acquired in the conventional walk analysis. The meaning of plantar slip measurement is explained and the plantar skin deformation is measured by the image processing using the developed measurement device. An increase of the plantar skin contact area by using ice immersion approach suggests the relation between plantar tactile stimulus and gait pattern.

This paper introduces a person following shopping cart robot. The developed prototype robot is using a shopping cart as the robot's main body and attaching a driving module. The aim of this study is to develop a porting system for large-sized facilities (i.e. shopping mall, airport, etc.), where the object item is autonomously carried by a person following ("after you") porting robot. We believe such study on porting robots is highly practical and necessary in today's aging society. By using the computer's signal output for autonomous run and the manually control function for risk aversion, a compatible control with automatically and manually control is realized. The structure and the principle of control system are described

This paper describes a method of measuring the incipient slip of the sole in order to clarify the relation between a plantar tactile sensibility and a stabilization mechanism for a slip. The incipient slip measuring device for walking analysis by using high-speed cameras and acceleration sensor is developed. The measurement device can be measured the plantar skin deformation during a walk. The incipient slip is confirmed by using the developed device.

In this paper, the role of the human being toe is investigated. The change of the sole pressure distribution using a sole pressure distribution measurement system and that of the toe pressure using a toe pressure measurement system at walk by the existence of a toe are investigated. In order to investigate the change of the whole walking, the affection of a walk posture using a 3D motion capture system by the existence of a toe is also measured. Furthermore, the change of the amount of energy consumption at walk by the existence of a toe is measured by a spirometry system. At last, we investigate the roles of the toe, such as, the efficiency of the big and 2nd toe, and the stabilization of the posture, by the synthesis of these results.