Tananun CHOTPRA-SERTKOON93WarasineeCHAISAN-GMONGKON9495AditKURNIAWANThe Status and Distribution of Green Peafowl (Pavo muticus) in Northern Thailand: Providing a Baseline for Community-Based Management(Project 2019)Development and Validation of Tuberculosis and Pneumonia Detection Algorithms for Chest X-Ray Images in Thai Population(Project 2020)Cognitive Radio Technique for 5G/6G Wireless Communications System(Project 2020)66The endangered Green Peafowl have declined dramatically across their range due to hunting and habitat degradation. In Thailand, there are two strongholds remain in the west and north. While the western stronghold has been continuously investigated and well protected, the northern stronghold is remained unknown. This work aim to investigating the species status within northern stronghold covered four protected areas and surrounding agriculture landscapes. Using Distance sampling over 54 transect, 2 km long, located in the interior, edge and agriculture landscape. From the survey, overall estimated is 15.82 calling males/km2 throughout study area. The general linear models showed that species distribution was affected positively by dry dipterocarp forest and negatively by human communities. While in transect survey the species was positively affected by ground vegetation cover and presence of fire and negatively affected by presence of human activities, domestic dog and cattle. Despite much promising research in the area of artificial intelligence for medical image diagnosis, as of yet there has been no large-scale validation study done in Thailand to confirm the accuracy and utility of such algorithms when applied to local Thai datasets. Here we present the development and testing of a deep learning algorithm for automated thoracic disease detection, utilizing 421,859 local chest radiographs. Our study shows that convolutional neural networks can achieve remarkable performance in detecting 7 common abnormality conditions on chest X-ray (CXR), and the incorporation of local images into the training set is key to the model's success. This paper presents a state-of-the-art model for CXR abnormality detection, reaching an average AUROC of 0.964. We also developed a high-performance Tuberculosis-focused model, with AUROC of 0.935. Our work emphasizes the importance of investing in dataset development and local research of medical diagnosis algorithms to ensure safe and efficient usage within the intended geographic region.The next evolving fifth generation (5G) wireless networks are envisioned to provide higher data rates, enhanced end-user quality-of-experience (QoE), reduced end-to-end latency, and lower energy consumption. This proposed research investigated and proposed two methods, which will enable high-capacity and spectrum-efficient access technology to support a larger number of users. The first method was to improve the rotate modulation scheeme using feedback channel measurement to correct the channel impairement, and to investigate the effect of feedback delay in practical implementation. The second method was to use low complexity Sparse Code Multiple Access (SCMA) system using Low Density Signature Orthogonal Frequeny Division Multiplexing (LDS-OFDM) system. The results show that both methods can produce improvemnet of bit error rate (BER) performance as a function of signal-to-noise ratio (SNR).
元のページ ../index.html#70