The Vietnam Skink ( Plestiodon tamdaoensis ) was described from Tam Dao National Park, Vietnam, in 1937, and the species is currently known only northern Hong Kong of China, whereas Devouring Forest Scincella devorator Yen Tu Nature Reserve, 2004, endemic to Vietnam. As a result our field surveys 2023 2024 Ba Vi we reported new distribution records, morphological data, natural history . In addition, used modeling predict potential these species. model showed that P. approximately 110,000 km...

In recent years, the binary coding of face image features, such as local patterns (LBP) and ternary (LTP) have become popular in recognition systems. These feature descriptors provide a simple powerful means for texture description. this paper, we present novel approach, which uses these to represent images, similarity feature-based selection classification algorithm improve rate. The is first divided into small regions from LBP LTP histograms are extracted concatenated single vector....

This paper addresses a new approach based on the Weber-face and singular value decomposition (SVD) methods to improve recognition accuracy for face system using local binary patterns ternary in an illumination variation environment. The images are first extracted illumination-invariant components by method. Secondly, SVD is applied encoded images. Next, training features or patterns. Finally, classification phase, matrix of test image combined with those adjust before classified. performed...

We propose a local texture descriptor, intensity area descriptor (LIAD), which is applied for human facial recognition in ideal and noisy conditions. Each image divided into small regions from LIAD histograms are extracted concatenated single feature vector to represent the image. The performed using nearest neighbor classifier with histogram intersection chi-square statistics as dissimilarity measures. Experiments were conducted ORL database of faces (Olivetti Research Laboratory,...

In this paper, we propose the effective similarity feature-based selection and classification algorithm to select features on training images classify face in recognition system. The experiments are conducted ORL Database of Faces, which consists 400 40 individuals. Two systems, one based histogram-based feature, other feature is mean pixel values window with size 4×4 (M4×4), developed. Euclidean distance Manhattan taken as metrics for method. results indicated that proposed algorithms not...

Interpretable Machine Learning faces a recurring challenge of explaining the predictions made by opaque classifiers such as ensemble models, kernel methods, or neural networks in terms that are understandable to humans. When model is viewed black box, objective identify small set features jointly determine box response with minimal error. However, finding model-agnostic explanations computationally demanding, problem intractable even for binary classifiers. In this paper, task framed...

Illumination variation is among the several bottlenecks in a face recognition system because it can greatly affect appearance of image, which causes reduction matching performance system. Using illumination preprocessing methods an effective approach to overcome this problem. Despite achievements made, however, each method still has its own demerits. In paper, efficient representation insensitive varying based on combination Weber-face (WF) and difference Gaussians (DoG) proposed for human...

Computed Tomography (CT) technology has been widely used for detecting and diagnosing lung disease. To improve the visibility of essential features in a CT image, reducing noise blur, sharpening features, increasing contrast are necessary. In this study, we propose method to these gain better characteristics region interest images right diagnosis. Processing proposed consists non-local means filter removing noise, unsharp masking deblurring finally singular value decomposition enhancing...

We present, in this short paper, a model of artificial brain based on the Software-Hardware integration "1 + 1 = 1" philosophy framework using machine learning and multiprocessor system chip, SoC.Its virtual experiences are generated by deep process with random changing structure net neural network, NoNN, Monte Carlo method.It ensures creative property human cognitive processing possibility "humanmachine" integration/"Human brain-Artificial Brain" integration, which should be applied various...

Model-based estimation often uses impact factors and historical data to predict the effort of new projects. Estimation accuracy this approach is highly dependent on how well are selected. This paper comparatively assesses six methods for prune parameters models, including Stepwise regression, Lasso, constrained GRASP, Tabu search, PCA. Four sets were used evaluation, showing that varies among but no method consistently outperforms rest. regression prunes model most while it does not...

Abstract4Local phase quantization (LPQ) descriptor, first introduced by Ojansivu and Heikkila (2008), has successfully been applied in face recognition systems.In this paper, we combine local intensity area descriptor (LIAD), which was Tran (2017), with LPQ to develop robust systems using descriptor.Face images were encoded LIAD as a noise dimensionality reduction step.After that, the resulting presented through feature extraction step.A nearest neighbor method chi-square measure is used...

We open-source a state-of-the-art 4B-parameter generative model series for Vietnamese, which includes the base pre-trained monolingual PhoGPT-4B and its chat variant, PhoGPT-4B-Chat. The model, PhoGPT-4B, with exactly 3.7B parameters, is from scratch on Vietnamese corpus of 102B tokens, an 8192 context length, employing vocabulary 20480 token types. PhoGPT-4B-Chat, modeling output obtained by fine-tuning dataset 70K instructional prompts their responses, along additional 290K conversations....

The research study of detecting multiple intents and filling slots is becoming more popular because its relevance to complicated real-world situations. Recent advanced approaches, which are joint models based on graphs, might still face two potential issues: (i) the uncertainty introduced by constructing graphs preliminary slots, may transfer intent-slot correlation information incorrect label node destinations, (ii) direct incorporation intent labels for each token w.r.t. token-level voting...

Viewing improvement is an important step in lung computed tomography (CT) image analysis. It assists to make diagnosis more accurate. In this paper, a new viewing method for CT images introduced. The proposed consists of two stages. the first stage, was deblurred and sharpened using unsharp masking method. second pre-processed enhanced by modified method, thresholded local intensity area descriptor. experiments were conducted on from LIDC-IDRI database. For evaluation, our evaluated...