Бидний тухай
Багш ажилтан
Хиймэл оюун ухаан хурдацтайгаар хөгжиж түүнийг дагаад хүмүүс, тэр дундаа дэлхийн олон хүн ам хэрэглэдэг Англи гэх мэт түгээмэл шиглагддаг хэлээр хүн машинтай харьцах боломж хангалттай түвшинд хөгжөөд байгаа бол Монгол хэлний хувьд бусад хэлтэй харьцуулахад энэ нь хязгаарлагдмал байна. Тиймээс бид энэхүү ажлын хүрээнд монгол хэлний төхөөрөмжтэй харилцах эхний шат болох Монгол хэлээр ярьсан яриаг илтгэгчээс хамааралгүй текстэд хөрвүүлэх, түүнийг өөр нэгэн зорилтот хоолойд хувиргах ажлыг Whisper болон MMS моделийн тусламжтай хийсэн. Энэхүү судалгааны ажлын үр дүнд суурилаад Монгол хэлээр ярьж буй яриаг текстэнд хөрвүүлэх программ, цаашлаад Монгол хэлдээр ажиллах боломжтой харилцан ярианы системийг хөгжүүлэх боломжтой.
Spiking Neural Networks are a type of artificial neural network that mimics the way biological neural networks in the brain process information. Spiking neural networks form the foundation of the brain’s efficient information processing. While we don’t fully understand how these networks calculate, recent optimization techniques allow us to create increasingly complex functional spiking neural networks in a simulated environment. These methods promise to develop more efficient computing hardware and explore new possibilities in understanding brain circuit function. It is essential to have objective methods to compare their performance to speed up the development of such techniques. However, there are currently no widely accepted means of comparing the computational performance of spiking neural networks. We have introduced a new spike-based classification dataset that can be widely used to evaluate software performance and neuromorphic hardware implementations of spiking neural networks to address this issue. To achieve this, we have created a general procedure for converting audio signals into spiking neural network activity, drawing inspiration from neurophysiology. We created the Monnum digit dataset specifically for this study. Within the range of this research, We implemented a digit recognition system from 1 to 10 spoken in the Mongolian language for the Spike neural network. The last is data for training and testing, which was prepared in HDF5 format extension and then trained in the SNN network.
The tradition of Mongolian poems and poetry has a very ancient history, and long song poems are an expression of the Mongolian people's mind, especially Ayzam long songs perfectly contain the philosophical characteristics of the world, worldly phenomena, and the nature of human life, while suman In addition to common themes such as common people's life customs, animal husbandry culture, respect for ancestors and father-in-law, worshiping mountains and water, the long songs have a wide range of content that allows you to recognize the rise and fall of Mongolian history. But besreg songs, like suman songs, are dominated by respect, love, customs, as well as message and education. It can be seen from these that the history of Mongolians in any era is steeped in long songs and poems. Mongolian long song is one of the major genres of Mongolian folklore that has been widespread among Mongolians since ancient times. The emergence of the long song is based on the nature of work or the song of work. As the long song spread among the ancient nomads of Asia, it developed into the feast, festival, and ceremonial song of the Mongols in the 12th and 13th centuries, and then it became a form of long song and was passed down through many generations. The study of long songs has been studied at the appropriate level in terms of melody and rhyme. However, there are very few researches at the level of poems and words using modern experimental research methods. In this work, we have analyzed the system of the long song poem and the features of the words on the example of the long song poem of the Central Khalkh, while using some modern information technology experiments and research methods, we have tested the poem and the words and obtained appropriate results. Mongolian folk songs are very rich, so we have specially analyzed and studied the poems of Central Khalkh folk songs. In terms of poetics, the rhyming system is clearly reflected in Mongolian folklore genres such as proverbs, three worlds, etc., which are spoken in a uniform rhythm, but it is not always observed in long song poems. But syllable-accented and accented verse systems are common in long lyric poems. However, in our experiments with long song poems, it was seen that the melodic accent was placed on the second word of the poem, and it was sung in a melodious retro style. As an example, experiments were conducted on two songs that are representative of the central pitch length. In doing so, Praat (Praat) was used to record the sound image and adjust the size of the lyrics to the time of singing. The Praat program is used by scientists from the "Language Experimental Research Center" of the Ministry of Education of the National University of Moscow and the "Machine Intelligence Laboratory" of the National Technical University of Moscow State University and have produced results. For us, it was the first time we experimented with the melody of a long song. Some statistical and quantitative analyzes were performed on the collected song poems. In doing so, I downloaded the open data processing program "MAXQDA" and analyzed the commonly used words in long song poems. In this way, the results that clearly express the characteristics of aizam, suman, and short-length songs are shown.
The worldwide adoption of telehealth services may benefit people who otherwise would not be able to access mental health support. In this paper, we present a novel algorithm to obtain reliable pulse and respiration signals from non-contact facial image sequence analysis. The proposed algorithm involved a skin pixel extraction method in the image processing part and signal reconstruction using the spectral information of RGB signal in the signal processing part. The algorithm was tested on 15 healthy subjects in a laboratory setting. The results show that the proposed algorithm can accurately monitor respiration rate(RR), pulse rate (PR), and pulse rate variability (PRV) in rest conditions.
Robotic arms are widely used for many applications in various fields such as industry, health care, and military. Most robotic arms are programmed to perform specific tasks. However, controlling the robotic arm to move in any position is not easy and requires experience. In this study, we used Leap motion sensor to detect and track human hand movement and trajectory. We developed a mathematical model that converts human hand coordinates that are acquired from Leap motion sensor to 6-axis robotic arm control coordinates using Python language. To evaluate performance, we compared human hand coordinates with controlled 6-axis robotic arm coordinates. As a result, the 6-axis robotic arm followed the human hand trajectory successfully. Therefore, using Leap motion sensor and the proposed mathematical model, 6-axis robotic arm can follow the human hand movement and trajectory accurately. Based on the result, in future, we can develop the robotic arm chess player that plays chess automatically without human help.
Дуу хоолойг хувиргах гэдэг нь бодит цаг хугацааны хувьд хүний яриаг өөр хүн ярьж байгаа юм шиг дуу хоолойг нь хувиргах буюу өөрчлөхөд ашигладаг техник юм. Хүний дуу хоолой хувиргалтын алгоритмыг ISP болон зар сурталчилгааны компаниуд түгээмэл хэрэглэдэг. Гэхдээ ихэвчлэн дуу хоолой хувиргалтын алгоритмын хэрэгжүүлэлтийг бэлэн платформ буюу нээлттэй-эх код ашиглан хийдэг бөгөөд энэ нь дохиог өөрсдийнхөө хүссэн дуу хоолойруу хөрвүүлэх гэх мэт боломжуудыг хязгаарладаг. Иймээс бид энэхүү судалгааны ажлаар Питч шилжүүлэлтийн алгоритмыг МАТЛАБ програм болон C++ хэл дээр хэрэгжүүлсэн. Эхлээд хүний ярианы дохиог тодорхой хэмжээтэй фрейм болгон хуваасан бөгөөд нэг фрейм нь дохионы питч буюу үндсэн давтамж юм. Фрейм болгон хуваахын тулд бид тухайн хүний ярианы дохионоос питч илрүүлэх процессийг хийнэ. Өөрөөр хэлбэл тухайн хүний ярианы давтамжийн хүрээг олно. Хэрэгжүүлэлтийг монгол хүний дуу хоолой ашиглан хийсэн бөгөөд питч үнэлэх, питч тэмдэглэх, Pitch Synchronous Over Lap Add алгоритм гэсэн гурван алхмын дагуу хийсэн. Бидний энэхүү систем нь дурын хүний дуу хоолойн дохиог далайц болон хугацааны хувьд өөрчлөх замаар хувиргаж хувиргасан дуу хоолойг гаргадаг. Мөн энэхүү алгоритмын тусламжтайгаар хүний ярианы дохиог хувиргасан хэдий ч дууны чанарыг сайжруулах шаардлага гарсан учраас далайцын утгуудыг сэргээх алгоритмыг мөн хэрэгжүүлсэн.
Heart rate is one of the main vital signs that reflect human health conditions. There are various heart rate measuring devices have been developed. But high-quality devices are expensive and low-cost devices are less accurate. This paper proposes to develop a cost-effective, portable, and accurate heart rate measuring device using an inexpensive photoplethysmography sensor. The proposed device measures heart rate in 10 seconds and the numerical value is shown on the OLED display. Heart rate signal processing and computation are conducted using an Arduino Nano microcontroller. We enrolled 64 subjects (age 6-66 years, 33 males, 31 females) to evaluate the accuracy of the proposed device in this study. Subjects were asked to sit steady and heart rates were measured by the proposed device from the fingertip and standard pulse oximetry simultaneously. Heart rates from the proposed device are compared with the reference measurement using Pearson correlation, student t-test, and Bland-Altman analysis. The correlation coefficient was r=0.96 with p<0.0001 and Bland-Altman analysis revealed the proposed device and reference values had good agreement with a mean difference of -1.6 bpm. The proposed device appears promising for heart rate measurements and the design concept is portable, lightweight, and user-friendly. Also using a photoplethysmography sensor the device could cost 2-5 times cheaper than other devices on market.
The primary cause of death among children under age 5 years is acute respiratory infection, such as pneumonia. Detection of infection at the earliest point of contagion is necessary, to reduce morbidity and prevent infectious disease epidemics; therefore, identifying abnormal vital signs is essential. For early detection of pediatric infections, we developed a low-cost, portable, rapid screening system of pediatric infection. The system simultaneously measures three vital signs: heart rate (HR), respiration rate (RR), and body temperature (Temp) within 10 seconds using a pulse sensor, Doppler radar, and an infrared thermopile. Vital sign signal processing and computation are conducted using an Arduino Nano microprocessor, enabling the small, lightweight, and portable design of this system. Moreover, the cost-effectiveness of the system facilitates system applications in developing countries, which have the highest levels of pediatric mortality. We conducted trial measurement in Bayangol Health Center, Ulaanbaatar, Mongolia in 2019. A total of 50 children (age 1-14 years, 26 boys/24 girls) were enrolled in this study. Bland-Altman plot and Pearson correlation methods were used to evaluate the accuracy of the proposed system. The correlation coefficients were calculated as HR: r=0.92, RR: r=0.8, and Temp: r=0.82, with p<; 0.01. The system appears promising for rapid and convenient detection of infection in children.
Үр дүнд суурилсан, оюутан төвтэй сургалтад шилжих шаардлагатай тулгараад буй өнөө үед оюутныг ойлгож, нөхцөл байдлыг нь мэдрэх, бодитоор үнэлэх асуудал сурах явцыг сайжруулах, шалгалтын арга хэлбэрийг илүү боловсронгуй болгох ач холбогдолтой юм. Оюутан өөрийн мэдлэг, шалгалтын бэлтгэл болон зан араншингаасаа хамааран аман шалгалтын явцад сандрах, баярлах, түгших хэлбэрээр сэтгэл нь хөдөлж улмаар энэ нь бодит үнэлгээнд нөлөөлөх талтай. Энэхүү ажлаар оюутны сэтгэл хөдлөл болон үнэлгээний хамаарлыг нөхцөл байдал, асуултын төрлөөс хамааруулан судаллаа. Сэтгэл хөдлөлийг тухайн оюутны арьсны цахилгаан дамжуулал, хөмсөг болон хацрын булчингийн цахилгаан идэвхжилийн утгаар тодорхойлсон. Туршилтыг хоёр үе шаттай явууллаа. 28 оюутныг мэдрүүлтэй болон ердийн гэсэн хоёр бүлэгт хуваан ижил агуулгаар аман шалгалт авсан эхний туршилтаар мэдрүүл зүүх нь оюутны сэтгэл хөдлөл, хариулсан үр дүнд мэдэгдэхүйц нөлөө үзүүлэхгүй гэж дүгнэж болохоор байна. Хоёр дахь туршилтад мэргэжлийн суурь болон мэргэшүүлэх хичээл сонгон тус бүр 10 оюутнаас энгийн тав, шалгалтын таван асуулт асууж сэтгэл хөдлөлийн өөрчлөлтийг судлав. Туршилтаар оюутан энгийн асуултад хариулахдаа сэтгэл хөдлөлийн төлөв нь бага өөрчлөгдөж байсан бол шалгалтын асуултад бүх оюутан илүү төвлөрснөөс гадна ахлах ангийн оюутнууд жигд бага зэрэг сөрөг сэтгэл хөдлөл үзүүлсэн бол бага ангийн оюутнууд эерэг сэтгэл хөдлөл илүүтэй харуулсан. Энэ судалгааг цаашид өргөжүүлэн төрөл бүрийн мэргэжил, нөхцөл байдлаар өргөтгөснөөр шалгалтын арга хэлбэрийг шинэчлэх, оюутны сэтгэл хөдлөл өөрийгөө илэрхийлэхэд нөлөөлдөг эсэхийг тодорхойлох, улмаар онлайн сургалтын шалгалтыг илүү баталгаатай болгох суурь нь болно.
Үр дүнд суурилсан, оюутан төвтэй сургалтад шилжих шаардлагатай тулгараад буй өнөө үед оюутныг ойлгож, нөхцөл байдлыг нь мэдрэх, бодитоор үнэлэх асуудал сурах явцыг сайжруулах, шалгалтын арга хэлбэрийг илүү боловсронгуй болгох ач холбогдолтой юм. Оюутан өөрийн мэдлэг, шалгалтын бэлтгэл болон зан араншингаасаа хамааран аман шалгалтын явцад сандрах, баярлах, түгших хэлбэрээр сэтгэл нь хөдөлж улмаар энэ нь бодит үнэлгээнд нөлөөлөх талтай. Энэхүү ажлаар оюутны сэтгэл хөдлөл болон үнэлгээний хамаарлыг нөхцөл байдал, асуултын төрлөөс хамааруулан судаллаа. Сэтгэл хөдлөлийг тухайн оюутны арьсны цахилгаан дамжуулал, хөмсөг болон хацрын булчингийн цахилгаан идэвхжилийн утгаар тодорхойлсон. Туршилтыг хоёр үе шаттай явууллаа. 28 оюутныг мэдрүүлтэй болон ердийн гэсэн хоёр бүлэгт хуваан ижил агуулгаар аман шалгалт авсан эхний туршилтаар мэдрүүл зүүх нь оюутны сэтгэл хөдлөл, хариулсан үр дүнд мэдэгдэхүйц нөлөө үзүүлэхгүй гэж дүгнэж болохоор байна. Хоёр дахь туршилтад мэргэжлийн суурь болон мэргэшүүлэх хичээл сонгон тус бүр 10 оюутнаас энгийн тав, шалгалтын таван асуулт асууж сэтгэл хөдлөлийн өөрчлөлтийг судлав. Туршилтаар оюутан энгийн асуултад хариулахдаа сэтгэл хөдлөлийн төлөв нь бага өөрчлөгдөж байсан бол шалгалтын асуултад бүх оюутан илүү төвлөрснөөс гадна ахлах ангийн оюутнууд жигд бага зэрэг сөрөг сэтгэл хөдлөл үзүүлсэн бол бага ангийн оюутнууд эерэг сэтгэл хөдлөл илүүтэй харуулсан. Энэ судалгааг цаашид өргөжүүлэн төрөл бүрийн мэргэжил, нөхцөл байдлаар өргөтгөснөөр шалгалтын арга хэлбэрийг шинэчлэх, оюутны сэтгэл хөдлөл өөрийгөө илэрхийлэхэд нөлөөлдөг эсэхийг тодорхойлох, улмаар онлайн сургалтын шалгалтыг илүү баталгаатай болгох суурь нь болно.
Despite a reduction in pneumonia-related mortality, pneumonia remains a leading cause of death among children aged 0–5 years. Most of these deaths occur in developing countries. However, more than half of pneumonia-related deaths are preventable with improved facilities and health strategies. Early rapid diagnosis is important to decrease pneumonia mortality. We developed a portable, cost-effective, rapid pneumonia screening system using a random tree algorithm to support early detection of pneumonia in children. We enrolled 105 participants: 57 patients aged 1–13 years (33 boys, 24 girls) diagnosed with pneumonia by chest radiograph and 48 normal volunteers aged 2–14 years (25 boys, 22 girls). We conducted a clinical trial in the Bayangol District Geriatric and Pediatric Hospital, Ulaanbaatar, Mongolia from January 12–19, 2019. Our screening system measured heart rate, respiration rate and skin temperature within 10 seconds and used a random tree algorithm to distinguish patients with pneumonia and normal volunteers. The system uses a photosensor, Doppler radar, and infrared thermophile to determine vital signs and an Arduino Nano microprocessor to perform computations. Paired t-tests were used to compare vital signs between patients with pneumonia and normal volunteers. The random tree algorithm achieved sensitivity of 96.5%, specificity of 81.3%, positive predictive value of 85.9%, and negative predictive value of 95.1%. The paired t-tests showed strong statistically significant differences in all three vital signs between patients with pneumonia and normal volunteers. Our random tree algorithm-based screening system offers an effective, rapid, and convenient tool for early detection of pneumonia in children. Its cost-effectiveness enables application in low-income countries. The system measures multiple vital signs simultaneously within 10 seconds, which may be useful for initial physical examinations in pediatric hospitals.
We previously reported in Journal of Infection an infection screening method using a combination of neural network and k-means clustering algorithm.1 After a pandemic of a novel influenza A/H1N1pdm09 in 2009, we started to develop vital sign-based infection screening systems and conducted several preliminary infection screenings in Japan, Mongolia and Vietnam.2–4 Although the previous method achieved up to 98% sensitivity and 96% negative predictive value (NPV), this study was limited to patients with comparatively uniform backgrounds, i.e., hospitalized influenza patients who were soldiers of the Japan Self-Defense Forces.
Fever is one significant sign of infection. Hence, infrared thermography systems are important for detecting infected suspects in public places. Reliable temperature measurements by thermography are influenced by several factors, including environmental conditions. This paper proposes a linear regression analysis-based facial temperature optimization method to improve the accuracy of multiple vital signs-based infection screening at various ambient temperatures. To obtain the relationship between ambient temperature and thermography measurements, 20 instances of axillary temperature, thermography measurements of facial temperature, and five different ambient temperature values at the time of measurement were used as a training set for a linear regression model. Temperatures from a total of 30 subjects were recalculated by the model. The screening system was evaluated using the temperature both before and after optimization to demonstrate the accuracy of the optimization method. A k-nearest neighbor algorithm was used to classify potentially infected patients from healthy subjects. Although the system has already been evaluated in restricted environmental conditions, this is the first time it was tested in Ulaanbaatar, Mongolia. The results show that the Pearson's correlation coefficient between optimum and axillary temperatures increased to r = 0.82. Paired t-tests revealed that the optimized temperature became statistically highly significant (p<;0.001) for differentiating potentially infected patients from healthy subjects. Finally, the system achieved a sensitivity score of 91% and a negative predictive value of 92%. These values are higher than those obtained without temperature optimization. The proposed optimization method is feasible and can notably improve screening performance.
Over 350 million people across the world suffer from major depressive disorder (MDD). More than 10% of MDD patients have suicide intent, while it has been reported that more than 40% patients did not consult their doctors for MDD. In order to increase consultation rate of potential MDD patients, we developed a novel MDD screening system which can be used at home without help of health-care professionals. Using a fingertip photoplethysmograph (PPG) sensor as a substitute of electrocardiograph (ECG), the system discriminates MDD patients from healthy subjects using autonomic nerve transient responses induced by a mental task (random number generation) via logistic regression analysis. The nine logistic regression variables are averages of heart rate (HR), high frequency (HF) component of heart rate variability (HRV), and the low frequency (LF)/HF ratio of HRV before, during, and after the mental task. We conducted a clinical test of the proposed system. Participants were 6 MDD patients (4 females and 2 males, aged 23–60 years) from Shizuoka Saiseikai General Hospital psychiatry outpatient unit and 14 healthy volunteers from University of Electro-Communications (6 females and 8 males, aged 21–63 years). The average PPG- and ECG (as a reference)-derived HR, HF and LF/HF were significantly correlated with each other (HR; r = 1.00, p < 0.0001, HF; r = 0.98, p < 0.0001, LF/HF; r = 0.98, p < 0.0001). Leave-one-out cross validation (LOOCV) revealed 83% sensitivity and 93% specificity. The proposed system appears promising for future MDD self-screening at home and are expected to encourage psychiatric visits for potential MDD patients.
After an outbreak of severe acute respiratory syndrome (SARS) in 2003, infrared thermography (IRT) systems were used as the border control strategies for fever screening of passengers at most major international airports. In this study, we proposed a visible/thermal image processing approach using only a CMOS camera-equipped IRT, thereby remote sensing of multiple vital signs for the rapid and accurate screening. Due to infection screening system expensive and required large-scale system structure, the system still has not reached yet the stage of widespread use. In this regard, we focused on minimum the hardware requirements to achieve a system that more suitable for real-world settings. The system displays the results via the logistic regression discriminant function derived from the heart and respiration rate, and body temperature within 10 seconds. The system achieved a sensitivity and specificity of 87.5% and 91.7%, which is notably higher compared to conventional fever based screening.
Background: Infrared thermography (IRT) is used to screen febrile passengers at international airports, but it suffers from low sensitivity. This study explored the application of a combined visible and thermal image processing approach that uses a CMOS camera equipped with IRT to remotely sense multiple vital signs and screen patients with suspected infectious diseases. Methods: An IRT system that produced visible and thermal images was used for image acquisition. The subjects’ respiration rates were measured by monitoring temperature changes around the nasal areas on thermal images; facial skin temperatures were measured simultaneously. Facial blood circulation causes tiny color changes in visible facial images that enable the determination of the heart rate. A logistic regression discriminant function predicted the likelihood of infection within 10 s, based on the measured vital signs. Sixteen patients with an influenza-like illness and 22 control subjects participated in a clinical test at a clinic in Fukushima, Japan. Results: The vital-sign-based IRT screening system had a sensitivity of 87.5% and a negative predictive value of 91.7%; these values are higher than those of conventional fever-based screening approaches. Conclusions: Multiple vital-sign-based screening efficiently detected patients with suspected infectious diseases. It offers a promising alternative to conventional fever-based screening.
Noncontact monitoring of respiration rate plays an important role on diagnosis of respiratory disorders, not only in hospital settings, but also in home healthcare. Piezoelectric pressure sensor (PPS) and Doppler radar have been developed for noncontact measurement of tiny body movements caused by spontaneous respiration. Such noncontact biosensors, unlike contact-type sensors, suffer from sensitivity to random motion artifacts. A hybrid PPS and Doppler radar sensor was considered as one of the solutions for addressing the issue of motion artifact. Hence, prior to the coupling of respiration signals from PPS and radar, a comparative study was conducted between these two sensors on five subjects with supine and lateral positions in a lab environment. We compared the respiration rates detected by noncontact sensors i.e., PPS and Doppler radar with reference contacttype respiratory effort belt, respectively. We observed that Doppler radar performed slightly better than PPS on supine position, whereas PPS was better on lateral position. This result implies that it is preferable to combine the two sensors in order to provide more accurate respiratory monitoring.
Pandemics of Severe Acute Respiratory Syndrome (SARS) in 2002 and H1N1 Flu (Swine Flu) in 2009 indicated repeated epidemics in various regions around the world. In order to prevent such spread of diseases, most effective approach is to detect infected individuals in their early stages. We have developed a novel infection screening system which detects the possibility of infection from measured vital signs, such as, heart rate, respiratory rate, and facial temperature within ten seconds. This study aims to evaluate the infection screening system for its accuracy when used in Mongolia, especially for tuberculosis patients and feverous ambulant patients. At present, we modified an infection screening system (KAZEKAMO) in National University of Mongolia for the clinical study at the National Centre. Prior to the clinical study in Mongolia, we conducted a clinical trial in Japan. Within ten seconds, the infection screening system with classification algorithm using a neural network discriminates potentially infected individuals from normal subjects using heart rates, respiratory rates, and facial temperatures measured by a pulse sensor, a compact radar, and thermography, respectively. The system was tested on 57 seasonal influenza patients after antiviral agent uptake (35.70C ≤ body temperature0C ≤ 38.30C ≤, 19-40 years) and 35 normal control subjects (35.50C ≤ body temperature0C ≤ 36.90C, 21-35 years) at the Japan Self-defense Forces Central Hospital. At the time of measurement, approximately half of the influenza patients have already became afebrile because of antiviral agents. The system achieved sensitivity of 98%, NPV of 82%, respectively. Our system appears promising for first order screening application of tuberculosis patients who sometimes develop only low-grade fever, because this system showed its efficacy for unfebrile patients.
