Search

Since SARM1 mutations have been identified in human neurological disease, SARM1 inhibition has become an attractive therapeutic strategy to preserve axons in a variety of disorders of the peripheral (PNS) and central nervous system (CNS). While SARM1 has been extensively studied in neurons, it remains unknown whether SARM1 is present and functional in myelinating glia? This is an important question to address. Firstly, to identify whether SARM1 dysfunction in other cell types in the nervous system may contribute to neuropathology in SARM1 dependent diseases? Secondly, to ascertain whether therapies altering SARM1 function may have unintended deleterious impacts on PNS or CNS myelination? Surprisingly, we find that oligodendrocytes express sarm1 mRNA in the zebrafish spinal cord and that SARM1 protein is readily detectable in rodent oligodendrocytes in vitro and in vivo. Furthermore, activation of endogenous SARM1 in cultured oligodendrocytes induces rapid cell death. In contrast, in peripheral glia, SARM1 protein is not detectable in Schwann cells and satellite glia in vivo and sarm1/Sarm1 mRNA is detected at very low levels in Schwann cells, in vivo, in zebrafish and mouse. Application of specific SARM1 activators to cultured mouse Schwann cells does not induce cell death and nicotinamide adenine dinucleotide (NAD) levels remain unaltered suggesting Schwann cells likely contain no functionally relevant levels of SARM1. Finally, we address the question of whether SARM1 is required for myelination or myelin maintenance. In the zebrafish and mouse PNS and CNS, we show that SARM1 is not required for initiation of myelination and myelin sheath maintenance is unaffected in the adult mouse nervous system. Thus, strategies to inhibit SARM1 function to treat neurological disease are unlikely to perturb myelination in humans.

Patients with Parkinson’s disease (PD) exhibit impaired dual-task (DT) performance. A recent meta-analysis confirmed that dual tasking severely affects walking performance in PD patients. However, one report indicated that a cycling DT paradigm has facilitative effects on cognition. We investigated the effects of dual tasking by using walking and cycling as motor tasks and revealed the clinical determinants associated with DT performance. Twenty-seven eligible participants were enrolled for clinical, cognitive-walking, and cognitive-cycling DT paradigm investigations. The mean age and age at onset of the patients were 59.87 ± 6.3 and 53.11 ± 8.4 years, respectively. Both the off- and on-state akinesia subscores were worse on the more-affected side than on the less-affected side. However, the DT effects on the cycling and gait outcomes on both the more-affected and the less-affected side showed no significant differences. The DT effect on the two motor tasks and cognitive performance during a concurrent walking task declined. Nevertheless, the DT effect on cognition improved during cycling. The present study also revealed that the levodopa equivalent daily dosage was highly associated with cognitive-cycling performance and that the akinesia subscore was the most relevant factor that contributed to cognitive-walking performance. In conclusion, DT facilitation or interference might be mediated by the type of motor task applied. The cognitive-cycling DT paradigm had a facilitative effect on cognition. Cycling exercise may diminish motor dysfunction has been investigated. We suggest that cognitive-cycling DT training is a potential adjuvant therapeutic strategy for patients with PD to promote motor and cognitive functions.

Alzheimer’s disease (AD) is a progressive neurodegenerative disease with memory loss and cognitive decline. Neurofibrillary tangles (NFTs) formed by hyperphosphorylated Tau protein are one of the pathological hallmarks of several neurodegenerative diseases including AD. Heat shock protein family B (small) member 1 (HSPB1) is a molecular chaperone that promotes the correct folding of other proteins in response to environmental stress. Nuclear factor erythroid 2-like 2 (NRF2), a redox-regulated transcription factor, is the master regulator of the cellular response to excess reactive oxygen species. Tropomyosin-related kinase B (TRKB) is a membrane-bound receptor that, upon binding brain-derived neurotrophic factor (BDNF), phosphorylates itself to initiate downstream signaling for neuronal survival and axonal growth. In this study, four natural flavones such as 7,8-dihydroxyflavone (7,8-DHF), wogonin, quercetin, and apigenin were evaluated for Tau aggregation inhibitory activity and neuroprotection in SH-SY5Y neuroblastoma. Among the tested flavones, 7,8-DHF, quercetin, and apigenin reduced Tau aggregation, oxidative stress, and caspase-1 activity as well as improved neurite outgrowth in SH-SY5Y cells expressing ΔK280 TauRD-DsRed folding reporter. Treatments with 7,8-DHF, quercetin, and apigenin rescued the reduced HSPB1 and NRF2 and activated TRKB-mediated extracellular signal-regulated kinase (ERK) signaling to upregulate cAMP-response element binding protein (CREB) and its downstream antiapoptotic BCL2 apoptosis regulator (BCL2). Knockdown of TRKB attenuated the neuroprotective effects of these three flavones. Our results suggest 7,8-DHF, quercetin, and apigenin targeting HSPB1, NRF2, and TRKB to reduce Tau aggregation and protect cells against Tau neurotoxicity and may provide new treatment strategies for AD.

This paper presents a deep learning based real-time two-stage gesture recognition system. In the first stage, the system deploys the YOLACT++ model to detect the human body mask. Then four ResNet18 models are used separately to predict the correctness of push-up, the number of push-ups, correctness of sitting posture, and the flexibility of the standing forward bend. The accuracy rates of these four ResNet18 models reached 97.9%, 97.5%, 95.1%, and 95.4%, respectively, using the test dataset collected from ten individuals.

To assist in the identification of possible dangerous situations in the elderly care situation, such as a fall. This study utilizes action recognition to detect and record elder daily movement. If there is any abnormality, the detection system will send out a warning for help. The accuracy is 87.5% for the elder action recognition that developed with CNN and LSTM.

Additional file 1: Supplementary Figure 1. Flow of study selection. NS indicates nephrotic syndrome; IS, ischemic stroke; ICH, intracranial hemorrhage. Supplementary Table 1. The numbers of missing laboratory data. Supplementary Table 2. Demographic features of nephrotic syndrome patients with stroke (additional data). Supplementary Table 3. Demographic features and clinical courses of nephrotic syndrome patients with ischemic stroke (additional data). Supplementary Table 4. Demographic features and clinical courses of nephrotic syndrome patients with intracranial hemorrhage (additional data). Supplementary Table 5. Cox regression analysis of patient survival during the 30-day period, intracranial hemorrhage vs. ischemic stroke (detailed items). Supplementary Table 6. Cox regression analysis of patient survival during the 30-day period in nephrotic syndrome patients with ischemic stroke (detailed items). Supplementary Table 7. Cox regression analysis of patient survival during the 30-day period in nephrotic syndrome patients with intracranial hemorrhage (detailed items).

Intelligent Customer Service is a very popular application at present. Using the previous user's Q & A records and the existing data in the database, combined with semantic analysis, users can quickly acquire what they want to know. The system takes the Taiwan Disease Control center data as an example, which is proved by experiments that the system can replace the traditional customer service work and can provide users with instant information about diseases.

The lack of snake knowledge not only makes humans easily fall into ophidiophobia, but also cause many snakes lose their lives. It is easy to encounter a snake in the countryside. It is necessary to identify whether it is a poisonous snake in the fastest time. In the city, although it is difficult to see the real snake, we have the need to understand the snake. In the future, if we could meet snakes in the mountains or villages, we can effectively activate the snake knowledge we have learned to avoid ourselves from dangerous situations. Based on the above reasons, this study considers a convenient, instant and interesting chatbot to assist the learning of snake knowledge. This study developed a Regional Snake Knowledge Learning System to assist in the promotion of snake-related knowledge, and effectively enhance learners' knowledge acquisition through the process. After the completion of the Snake Knowledge Learning System in this study, the effectiveness of this system on learner knowledge acquisition and memory retention results will be understood through the evaluation of the experimental group and the control group. In addition, this study also found that users have a high degree of satisfaction with the use of this system.

Many countries are facing a severe population ageing problem. The aging environment of the population, the trend of declining birthrate, the decline of family care function and the sense of loneliness in elderlies are impacting the lives of the elderlies. Even though there are many long-term care services at the national level it has been in implementation, there are still many restrictions on the recipients receiving service. Many elderly people in the healthy life stage are not eligible for service. However, based on the concept of “prevention is better than cure”, these healthy elders needs to be cared for in order to prevent the heavy resources and manpower generated in the future stage of treatment. The demand for long-term care is not only to meet their physiological needs, but also for their psychological needs. Therefore, this study hopes to use the standardization and information privacy protection for the elderly to care for huge amounts of health care data, and apply machine learning and statistical techniques, and deep learning to find out the hidden information of the elderly care, to find key decision points, and to establish alert and notification mechanism. Through this process, this study hopes to not only care for the physical and mental health of the elderly, but also to take care of the hearts and minds of the elderly and relatives around the elderly, and to establish a blessed smart community care sharing model.

The railway track fastener is the most important component of the rail fastening system. It can fasten the railway track onto the crosstie, and has great influence on train's stability and safety. In the early days, track patrol and maintenance were carried out on foot by track workers, and gradually adjusted to the use of track patrol vehicles, but still completely rely on manual visual inspection. However, the results of visual inspection may be limited by the vehicle ’s speed and inspection angle, or even long -term visual inspection may cause fatigue, and the problem of omission. In this study, we not only collect the related research results of track inspection, but also set up the image collecting device for track fastener including image recording equipment and lighting equipment. Yolo v3 is deployed and trained as the deep learning model, then the accuracy and recall rates of damaged fasteners are verified from the test dataset. In the experiments, we used the GoPro motion camera to capture images and record a total of 20 km track fastener images. The precision rate and recall rate of fasteners detection including defects were 89% and 95%, respectively, which can meet the demand of efficient inspection for railway track workers.

Traditional OCR system cannot handle scene texts. Text detection in natural scene images is much more difficult than the recognition of text in scanned document images because of its complicating background. Popular existing methods for characters segmentation are CTPN, EAST and PixelLink. However, they are not very capable of dealing with the small and densely character in large image, and connected characters. To cope with these problems, we adopted above popular character segmentation networks CTPN or EAST as the main structure and proposed using ResNet as the feature extraction network due to its excellent sensitivity of tiny features among those existing methods. The experimental result shows that the feature extraction network can affect the precision of locating text significantly. In the experiment with ICDAR dataset, the effect of deeper depth and larger width of ResNet on EAST is notable.

Gait disturbances in Parkinson's disease are commonly refractory to current treatment options and majorly impair patient's quality of life. Auditory cues facilitate gait and prevent motor blocks. We investigated how neural dynamics in the human subthalamic nucleus of Parkinsons's disease patients (14 male, 2 female) vary during stepping and whether rhythmic auditory cues enhance the observed modulation. Oscillations in the beta band were suppressed after ipsilateral heel strikes, when the contralateral foot had to be raised, and reappeared after contralateral heel strikes, when the contralateral foot rested on the floor. The timing of this 20-30 Hz beta modulation was clearly distinct between the left and right subthalamic nucleus, and was alternating within each stepping cycle. This modulation was similar, whether stepping movements were made while sitting, standing, or during gait, confirming the utility of the stepping in place paradigm. During stepping in place, beta modulation increased with auditory cues that assisted patients in timing their steps more regularly. Our results suggest a link between the degree of power modulation within high beta frequency bands and stepping performance. These findings raise the possibility that alternating deep brain stimulation patterns may be superior to constant stimulation for improving parkinsonian gait.SIGNIFICANCE STATEMENT Gait disturbances in Parkinson's disease majorly reduce patients' quality of life and are often refractory to current treatment options. We investigated how neural activity in the subthalamic nucleus of patients who received deep brain stimulation surgery covaries with the stepping cycle. 20-30 Hz beta activity was modulated relative to each step, alternating between the left and right STN. The stepping performance of patients improved when auditory cues were provided, which went along with enhanced beta modulation. This raises the possibility that alternating stimulation patterns may also enhance beta modulation and may be more beneficial for gait control than continuous stimulation, which needs to be tested in future studies.

Most traditional facial expression-recognition systems track facial components such as eyes, eyebrows, and mouth for feature extraction. Though some of these features can provide clues for expression recognition, other finer changes of the facial muscles can also be deployed for classifying various facial expressions. This study locates facial components by active shape model to extract seven dynamic face regions (frown, nose wrinkle, two nasolabial folds, two eyebrows, and mouth). Proposed semantic facial features could then be acquired using directional gradient operators like Gabor filters and Laplacian of Gaussian. A multi-class support vector machine (SVM) was trained to classify six facial expressions (neutral, happiness, surprise, anger, disgust, and fear). The popular Cohn---Kanade database was tested and the average recognition rate reached 94.7 %. Also, 20 persons were invited for on-line test and the recognition rate was about 93 % in a real-world environment. It demonstrated that the proposed semantic facial features could effectively represent changes between facial expressions. The time complexity could be lower than the other SVM based approaches due to the less number of deployed features.

Over the last few years, mobile devices have become part of our lives. More and more new applications are being developed due to the portability of mobile devices equipped with multiple sensors. In this paper, we proposed a simple and efficient system for gesture-based mouse emulation on mobile devices. The system uses only an RGB camera as its input, but is still capable of handling various lighting conditions. The mouse emulation is implemented using an input module, which is one of Linux kernel modules that allows user to create and handle virtual input devices from an application. Thus, the input events generated by our system are injected system-wide, which acts like a real mouse plugged into the device. In addition, the input model is implemented using thread to speed up. Experimental results show that our system can achieve real-time performance with low latency on a low-cost mobile device.

See Vidailhet et al. (doi:10.1093/brain/awx140) for a scientific commentary on this article.Misdiagnosis among tremor syndromes is common, and can impact on both clinical care and research. To date no validated neurophysiological technique is available that has proven to have good classification performance, and the diagnostic gold standard is the clinical evaluation made by a movement disorders expert. We present a robust new neurophysiological measure, the tremor stability index, which can discriminate Parkinson's disease tremor and essential tremor with high diagnostic accuracy. The tremor stability index is derived from kinematic measurements of tremulous activity. It was assessed in a test cohort comprising 16 rest tremor recordings in tremor-dominant Parkinson's disease and 20 postural tremor recordings in essential tremor, and validated on a second, independent cohort comprising a further 55 tremulous Parkinson's disease and essential tremor recordings. Clinical diagnosis was used as gold standard. One hundred seconds of tremor recording were selected for analysis in each patient. The classification accuracy of the new index was assessed by binary logistic regression and by receiver operating characteristic analysis. The diagnostic performance was examined by calculating the sensitivity, specificity, accuracy, likelihood ratio positive, likelihood ratio negative, area under the receiver operating characteristic curve, and by cross-validation. Tremor stability index with a cut-off of 1.05 gave good classification performance for Parkinson's disease tremor and essential tremor, in both test and validation datasets. Tremor stability index maximum sensitivity, specificity and accuracy were 95%, 95% and 92%, respectively. Receiver operating characteristic analysis showed an area under the curve of 0.916 (95% confidence interval 0.797-1.000) for the test dataset and a value of 0.855 (95% confidence interval 0.754-0.957) for the validation dataset. Classification accuracy proved independent of recording device and posture. The tremor stability index can aid in the differential diagnosis of the two most common tremor types. It has a high diagnostic accuracy, can be derived from short, cheap, widely available and non-invasive tremor recordings, and is independent of operator or postural context in its interpretation.

The general public spends a significant amount of time in front of digital signage seeking information from many venues such as exhibition halls and shopping centers. This is why advertisement purchasers believe that the number of passing viewers provides crucial information for their marketing strategies. In this paper, a real-time person counting/memorizing system is designed capable of distinguishing the gender of potential customers. An adaptive boosting (Adaboost) machine learning algorithm is used to detect human faces and utilize specific filtering criteria to eliminate useless data. For each detected person, face and torso information are recorded in a database for identification. The least recently used identification record will be deleted if the database is full. Gender classification is performed by support vector machine using hair ratios extracted from gender characterizing regions. Based on a variety of experiments, the accuracy of the proposed algorithm is higher than 90 % for person-counting and higher than 94 % for gender classification. Moreover, the execution speed on personal computers may reach 15---20 fps.

Due to the effect of lighting and complex background, most visual hand gesture recognition systems work only under restricted environments. Here, we propose a robust system which consists of three modules: digital zoom, adaptive skin detection, and hand gesture recognition. The first module detects user face and zooms in so that the face and upper torus take the central part of the image. The second module utilizes the detected user facial color information to detect the other skin color regions like hands. The last module is the most important part for doing both static and dynamic hand gesture recognition. The region of interest next to the detected user face is for fist/waving hand gesture recognition. To classify the dynamic hand gestures under complex background, motion history image and four groups of novel Haar-like features are investigated to classify the dynamic up, down, left, and right hand gestures. A simple efficient algorithm using Support Vector Machine is developed. These defined hand gestures are intuitive and easy for user to control most home appliances. Five users doing 50 dynamic hand gestures at near, medium, and far distances, respectively, were tested under complex environments. Experimental results showed that the accuracy was 95.37 % on average and the processing speed was 3.93 ms per frame. An application integrated with the developed hand gesture recognition was also given to demonstrate the feasibility of proposed system.

In this work, we propose a reverse-order scheduling (ROS) method for broadcasting a video over a single channel. The method firstly partitions the video into equal-length segments, then divides these segments into groups, and finally broadcasts the segments in the same group over the same subchannel in the reverse order of their indices. Further, we also provide a thorough analysis of the performance of the ROS method. The main contribution of the ROS method is to exhibit not only a shorter waiting time but also a smaller buffer requirement. Compared with the Alternative Broadcasting (AB), Hopping-Insertion (HI), and SingBroad methods, the proposed ROS method reduces the average waiting time by 38%~96%, 18%~63%, and 3%~18%, respectively, while the bandwidth of a single channel is capable of more than triple the consumption rate. Further, the ROS method reduces the buffer requirement by 30%~70% and 30%~55% as compared with the AB and SingBroad methods, respectively, while the buffer requirement of the HI method is not given.

This paper proposed five new types of facial features for face recognition. Ada-boost is used to detect face firstly. False detected faces are removed by dynamic background modeling and skin color detection. Skewed face is also calibrated to achieve higher accuracy. Based on Active Shape Modeling, the five new types of facial features including gradient histograms of facial components, vertical/horizontal projection of facial edge points, signature of facial components, multiple vertical/horizontal line segments within facial shape, and face template could be extracted. According to the classification capability, features are associated with different weights while during matching. Nearest neighbor classifier is deployed for face recognition by using the averaged of feature points of a person as the center. The size of database is 200 people which are selected from the face databases of MIT and ESSEX. Five images per person were used for training and 491 images were tested. The recognition rate was 98.3% and the processing speed reached 220ms per frame on a general personal computer.

In the real world, the viewers may want to have an ability to choose a desired video from some interested videos by scanning the first few minutes of these videos. Obviously, a broadcasting method, which allows subscribers buffering the data of the first few minutes of all popular videos, could provide this service without the startup latency. Fortunately, the capacity of harddisk or flash memory will increase and prices will become more affordable, enough to support the space needed to save such a large amount of data. In this work, we propose a zero-latency scheduling scheme (ZLSS) for broadcasting popular movies. The ZLSS method not only provides a zero-latency service for subscribers who buffering the data of the first few minutes of a video, but also provides a service with a shorter waiting time for subscribers who buffering nothing of the desired video. Compared with the previous method, the ZLSS method reduces the buffer requirement and the maximum waiting time by 10~50%.

DVB-H is a mobile digital video broadcasting technology based on DVB-T with the advantages of time-slicing and multiprotocol encapsulation?forward error correction (MPE-FEC) to reduce the power consumption and enhance the stability of receiver. The encapsulated data is compatible with the Internet Protocol which makes the integration of broadcast and telecommunication much easier. Therefore, digital content can be transmitted by adding encapsulation for each layer. In this paper, we developed a portable middleware for DVB-H clients. It could parse the Transport Stream according to the IP Data Casting (IPDC) protocol stack to achieve interactive multimedia services. Moreover, channel information contained in Electronic Service Guide (ESG) could also be decoded. Several applications were developed on our middleware to demonstrate its feasibility as an interactive multimedia platform.

Colorectal cancer (CRC) is the predominant gastrointestinal malignancy and constitutes a major medical and economic burden worldwide. A thorough understanding of the oncogenes or genes related to tumorigenesis is the key to developing successful therapeutic strategies. Molecular analysis of feces constitutes a potentially potent and noninvasive method for detection of CRC. Using nested reverse transcription-polymerase chain reaction (RT-PCR) and amplified restriction fragment length polymorphism analysis, sloughed cells from the entire length of the colon and rectum were analyzed for expression of activating K-ras codon 12 mutants, which are becoming attractive targets for antisense treatment. K-ras codon 12 mutant sequences were detected in feces of 5% (1/20) of healthy controls, in feces of 41% (12/29) of CRC patients, in 10% (3/29) of isolates of tissue complementary DNA (cDNA), and in 14% (4/29) of isolates of genomic DNA. Age of patient was significantly associated with K-ras codon 12 sequences in feces: Patients with wild-type K-ras codon 12 sequences were significantly younger than those with mutated forms of K-ras codon 12. Fecal ribonucleic acid (RNA) analysis was demonstrated to be a useful for diagnosis of CRC. This technique may be suitable for screening and determinatign the clinical significance of active mutations of the K-ras gene in feces and would possibly be useful for identificating patients that would benefit from antisense therapy.

A fuzzy ARTMAP system is a system for incremental supervised learning of recognition categories and multi-dimensional maps in response to an arbitrary sequence of analog or binary input vectors. Fuzzy ARTMAP systems have been benchmarked against a variety of machine learning, neural networks, and genetic algorithms with considerable success. Owing to many appealing properties, fuzzy ARTMAP systems provide a natural basis for many researchers. Many different approaches have been proposed to modify fuzzy ARTMAP systems. In this paper, we propose a new approach to modifying a fuzzy ARTMAP system. We refer to the new system as the modified and simplified fuzzy ARTMAP (MSFAM) system. The aims of MSFAM systems are not only to reduce the architectural redundancy of the fuzzy ARTMAP system, but also to make extracted rules more comprehensible and concise. Four data sets were used for demonstrating the performance of the proposed MSFAM system.

Most current license plate detection (LPD) systems target only the vehicles in a single-lane or at a short distance from the camera, i.e. each frame contains only one vehicle. Most of these systems utilizing edge features to detect license plates on distant multi-lane vehicles directly would fail because there are many edge features and the edges of the license plate would be interfered with by different types of vehicles and the background. In this paper, we propose a hierarchical approach that detects individual vehicles first, locates the region containing the license plate, and finally extracts the license plate. To verify the effectiveness of the proposed system, four-lane and two-lane video samples were used for testing. The car detection rates were almost 100% for both the two-lane and the four-lane videos. The LPD rate was 85% for the two-lane video and 92% for the four-lane video.

Due to vigorous development of pattern recognition, traditional human form filling tasks would be replaced by automated processes. However, these automation processes are often misused for illegal behavior such as spam e-mail or application for website account. In order to prevent website owner from suffering the attacks of automated program, this paper proposed an innovative image-based CAPTCHA for distinguishing human and computer by embedding versatile characters in the images. The proposed method makes the characters indiscernible by automated image analysis technologies like scale-invariant feature transform while human can easily distinguish the location of the embedded characters. Our designed mechanism was capable to elude such kind of attacks. In experiments, 15 users were invited to test the system and the success rate is 86%. If wrong operations like clicking out of text boxes were excluded, the success rate reached 95%. Compare the average logging time with reCAPTCHA and HELLO CAPTCHA, the proposed method is faster than these two methods by 32 seconds and 115 seconds, respectively.

Total indole-3-acetic acid (IAA) levels (free IAA plus IAA conjugates) in Doritis pulcherrima (Doritis pulcherrima Lindley, cv. S 84-335-2) leaves were threefold lower in plants exposed to 30 d of short-day (SD) conditions than in plants exposed to 30 d of long-day (LD) conditions. Free IAA levels were also significantly lower in SD-treated leaves than those in 0 day of photoperiodic treatment and 30 d of LD. Results indicate that plants exposed to 30 d of SD contained significantly higher levels of ester-IAA and reduced amide-IAA concentrations compared to those grown under LD conditions. A high level of ester-IAA and a reduction of amide-IAA in leaves may be related to increased SD floral initiation ability in D. pulcherrima. Key words: Doritis pulcherrima, floral initiation, indole-3-acetic acid, photoperiod

—As computers and networks have been developed vigorously, distance learning could be integrated with computer vision techniques for the purpose of better learning effects. In this paper, we developed a distance yoga learning system for people to learn/play through the internet. The main point of the interactive learning system essentially consists in that the gesture performed by player, segmented by computer vision techniques , should possess the same silhouette for a given yoga posture. For better accuracy, the learning score is calculated by matching the distance transformation of the player silhouette with stored standard yoga posture. In the experiments, 23 postures were defined and six persons were invited to do each posture three times. About 86% of the difference between computer scores and the scores given by a yoga teacher falls within -2.5~2.5.

Most traditional expression classification systems track facial component regions such as eyes, eyebrows, and mouth for feature extraction. This paper utilized facial components to locate dynamic facial textures such as frown lines, nose wrinkle patterns, and nasolabial folds to classify facial expressions. Adaboost using Haar-like feature and Active Shape Model (ASM) are adopted to accurately detect face and acquire important facial feature regions. Gabor filter and Laplacian of Gaussian are used to extract texture information in the acquired feature regions. These texture feature vectors represent the changes of facial texture from one expression to another expression. Support Vector Machine is deployed to classify the six facial expression types including neutral, happiness, surprise, anger, disgust, and fear. Cohn-Kanade database was used to test the feasibility of proposed method and the average recognition rate reached 91.7%.

We attempted to specifically quantify transcripts of faecal cytokeratin 19 (CK19) and ribosomal protein L19 (RPL19) RNA expression of colorectal cancer and clarify their correlation with clinicopathological parameters and survival in combination.Solid fecal samples were collected and preserved before any treatment. Levels of faecal CK19 and RPL19 mRNA were measured using quantitative real-time PCR. An expression level higher than median value was defined as positive.Between April 2001 and June 2007, 92 patients were recruited. The levels of both markers increased in a trend as stage. Young patients (67 years) were correlated with higher rate of CK19+ (p = 0.001), so were higher stages but with borderline significance (p = 0.051). CK19+ and RPL19+ were highly correlated mutually (p = 0.001). Neither CK19+ (p = 0.12) nor RPL19+ (p = 0.14) alone was a prognostic factor of disease-free interval. However, CK19+/RPL19+ was shown to be with worse prognosis (p = 0.037), but not an independent factor in multivariate analysis with stage.Both markers were significantly higher in the patients of metastatic disease. The use of two markers will recognize the high risk group better than the single marker usage, though not reaching independent status yet. Multi-target strategy assay is suggested for fecal RNA examination.

Man machine interface by video analysis becomes popular recently. The most frequently adopted body parts for computer interaction are face and hands. Therefore, it is a very important topic to accurately extract face and hand regions from a sequence of images in real time. In this paper, we propose an adaptive skin color model which is based on face detection. Skin colors were sampled from extracted face region where non-skin color pixels like eyebrow or eyeglasses could be excluded. Gaussian distributions of normalized RGB were then used to define the skin color model for the detected person. To demonstrate the robustness of proposed model, experiments under diversified lighting and background were tested. Traditional methods based on RGB, Normalized RGB, and YC(subscript b)C(subscript r) were all implemented for comparison. From experimental results, skin color pixels could be detected for each person. The accuracy rate is 95.73% on average and is superior to previously mentioned methods.

A real time V shape hand gesture recognition system by motion history template matching is proposed in this paper. Initially, user moves one of his/her hands in a specified region for template setup. Then, possible moving hand regions can be extracted by overlapping motion masks and skin color image. After noise removal of these regions, motion history template matching is used to recognize the V shaped hand gesture for shape toleration. The valley point of V is used to navigate the cursor and the left/right fingertip defines the left/right mouse button. In experiments, users can operate the windows system without physically contacting any equipment. The processing speed is more than 30 fps for images of size 320*240. The accuracy rate for the V shape hand gesture recognition is 96.92%. As to the recognition rates for left click, right click, and double click, we have 100%, 100%, and 91.53%, respectively.

Hand gesture recognition based man-machine interface is being developed vigorously in recent years. Due to the effect of lighting and complex background, most visual hand gesture recognition systems work only under restricted environment. An adaptive skin color model based on face detection is utilized to detect skin color regions like hands. To classify the dynamic hand gestures, we developed a simple and fast motion history image based method. Four groups of haar-like directional patterns were trained for the up, down, left, and right hand gestures classifiers. Together with fist hand and waving hand gestures, there were totally six hand gestures defined. In general, it is suitable to control most home appliances. Five persons doing 250 hand gestures at near, medium, and far distances in front of the web camera were tested. Experimental results show that the accuracy is 94.1% in average and the processing time is 3.81 ms per frame. These demonstrated the feasibility of the proposed system.

A model-guided structural matching method for recognizing handwritten off-line Chinese characters is presented. According to the stroke writing sequence, each character is described by an on-line model, which is a one-dimensional (1D) string consisting of a stroke sequence interleaved with relationships between two consecutive strokes. After an unknown input is thinned and line approximated, all possible strokes are extracted. A tree is created to represent all possible combinations of the extracted strokes which have the same stroke sequence as the one defined in the model. In the recognition process, we match the strokes with those defined in the on-line model. The matching process is formulated as a tree searching algorithm guided by the relationships in the model. The modelled character is taken to be a candidate if there is a feasible path which satisfies the relationships defined in the on-line model and the number of missing strokes in the input is less than a given threshold. The input is recognized as the character with the greatest number of matched strokes among all candidates. Experimental results on 300 frequently used characters in a database called CCL/HCCR1, which contains 5401 Chinese characters and about 250 variations for each one, show that the recognition rate is over 90%.

Power wastage is one of the most serious problems at present. The slogan ldquoturn off lights when not in userdquo has long been going around, but lights turned on in unoccupied rooms are frequently seen, particularly in public agencies or schools. The power line communication protocol frame work differs from ones used in the past such as X-10 [1], CEBus [3], LonWorks [6], etc. This system is intended for public spaces, where efficient management and control on lighting devices are implemented. Hopefully, simple circuit designs will help to achieve the purpose of power saving. By means of existing internal power line of buildings as communication media, module inductive circuits are compatible with current power switches. Through the system power lines and RF communication techniques, all school luminaries are managed by PDA remote monitoring method. Lights not in use are switched off to achieve power saving.

We designed a real time moving object extraction and behavior analysis system to detect if any abnormal behavior like intrusion, halt, and wall climbing happened. Much work in the area of motion based video object segmentation is being done in the pixel domain which exploits the visual attributes and motion information. Given most existing images and videos stored in the compressed form, the specific manipulation algorithms can be applied to the compressed streams without full decoding of the compressed images/videos. In this paper, we could produce motion masks from I and P frames in MPEG video. Then, motion vectors decoded directly from P frames could be utilized to do object extraction. By clustering motion vectors conforming to the same motion, we could extract the moving regions. If moving regions are connected closely and of the same direction, they are claimed to be of the same moving object. Following the motion vectors in consecutive P frames, we could track the moving objects. A confidence is defined from these matched blocks. Finally, a couple of criterions are defined for object behavior analysis whether they intrude a monitored area, halt on road, or climb wall. Experimental results are given to demonstrate the feasibility of our system.

In this paper, the camera is utilized to input characters as it becomes standard equipment in most IT products. The tip of forefinger acts like the pen tip is used to write characters in front of the camera. The finger tip is extracted and the tracked trail is recorded as the handwritten character. A character recognition system based on on-line character model is also developed to verify its practical usage. In experiments, each character is written five times by five individuals. The forefinger tip can be used as an input device and integrated with MS-Windows. The fingertip tracking accuracy is 98.58% and the processing speed could reach 13.2 fps. The character recognition rate is 91.11% if confusing character set is included. It demonstrated the feasibility of proposed finger tip extraction method and its application.

In this paper, we designed a simple and fast visual surveillance system to track human position and to determine if any abnormal behavior like wall climbing and falling happened. By taking both time and background difference into considerations, illumination effects could be greatly reduced while calculating motion masks. Refinements including holes filling, shadow removal, and noise reduction are done to obtain much more reliable motion masks. However, motion masks corresponding to occluded moving people, greater than a given width, are segmented recursively into smaller ones by bi-modal thresholding. Meanwhile, background could also be updated by the refined motion masks. Integrated location-based and weighted block-based matching is done for object tracking. A similarity is defined from these weighted matched block for object classification. Finally, a couple of criterions are defined to analyze whether objects stop, disappear, climb, or fall. Experimental results are given to demonstrate the robustness of our system.

This paper presents a reinforcement-learning approach to a navigation system which allows a goal-directed mobile robot to incrementally adapt to an unknown environment. Fuzzy rules which map current sensory inputs to appropriate actions are built through the reinforcement learning. Simulation results illustrate the performance of the proposed navigation system. In this paper, ACSNFIS is used as the main network architecture to implement the reinforcement-learning based navigation system.

This paper proposes a probabilistic approach to recognize handwritten Chinese characters. According to the stroke writing sequence, strokes and interleaved stroke relations are built manually as a 1-D string, called an on-line model, to describe a Chinese character. In an input character, strokes are first extracted by a tree searching method. The recognition problem is then formulated as an optimization matching problem in a multistage directed graph, where the number of stages is the length of the modelled stroke sequence. Nodes in a stage represent extracted strokes that have the same stroke type as defined in the on-line model and the link between two neighboring nodes corresponds to the relationship between the two extracted strokes. The probability that the extracted stroke belongs to the predefined stroke type is calculated from the stroke line segments, and the transition probability between two extracted strokes is the degree of satisfaction of the relationship defined in the on-line model. The Viterbi algorithm, which can handle stroke insertion, deletion, splitting, and merging, is applied to recover the sequence of strokes consisting of the unknown character. The similarity is defined to be the product of stroke probabilities and stroke transition probabilities in the stroke sequence. The unknown character is matched with all modelled characters and is recognized as the one with the highest similarity. Experiments with 540 characters uniformly selected from the database CCL/HCCR1 (250 variations/class) are conducted, and the recognition rate is about 92.8%, which proves the feasibility of the proposed recognition system.

In this paper, a new system for parcel map conversion and management is proposed. A scanner was used to input paper source maps into a computer. Text is separated from graphics by finding a connected component. Thinning and tracing are then executed by utilizing run-length codes for straight line extraction. A program is found if a closed loop is detected. All the extracted objects are stored into a spatial database. Attribute data such as land number, land use label, and land owner are saved into an attribute database. Topology representing the relations among objects is also generated and saved into the attribute database. We can query attribute data by spatial data and vice versa. A3 size parcel maps and Taipei administrative map are tested. The experimental results show that the proposed system is very effective and friendly. >