A5113416874- Automated Road and Building Extraction
- Image Processing and 3D Reconstruction
- Remote Sensing and LiDAR Applications
- Remote-Sensing Image Classification
- Advanced Algorithms and Applications
- Advanced Sensor and Control Systems
- Advanced Image and Video Retrieval Techniques
- Radiomics and Machine Learning in Medical Imaging
- Machine Learning in Materials Science
- Cell Image Analysis Techniques
- Advanced Image Fusion Techniques
- Advanced Neural Network Applications
- Generative Adversarial Networks and Image Synthesis
Xi’an University of Posts and Telecommunications
2024-2025
California University of Pennsylvania
2024
We introduce Baichuan-Omni-1.5, an omni-modal model that not only has understanding capabilities but also provides end-to-end audio generation capabilities. To achieve fluent and high-quality interaction across modalities without compromising the of any modality, we prioritized optimizing three key aspects. First, establish a comprehensive data cleaning synthesis pipeline for multimodal data, obtaining about 500B (text, audio, vision). Second, audio-tokenizer (Baichuan-Audio-Tokenizer) been...
The current generation of large language models (LLMs) is typically designed for broad, general-purpose applications, while domain-specific LLMs, especially in vertical fields like medicine, remain relatively scarce. In particular, the development highly efficient and practical LLMs medical domain challenging due to complexity knowledge limited availability high-quality data. To bridge this gap, we introduce Baichuan-M1, a series specifically optimized applications. Unlike traditional...
Nowadays, object detection algorithms are widely used in various scenarios. However, there further small requirements some special Due to the problems related objects, such as their less available features, unbalanced samples, higher positioning accuracy requirements, and fewer data sets, a algorithm is more complex than general algorithm. The effect of model for objects not ideal. Therefore, this paper takes YOLOXs benchmark network enhances feature information on by improving network’s...
This paper reports the application of deep learning techniques in bright-field transmission electron microscopy image segmentation and tracking, order to understand dynamic changes nanoparticles during high-temperature sintering. Four state-of-the-art models, YOLO v8n-seg, Swin- UNet, VMamba , EfficientSAM -tiny, were used their performances compared. The results show that -tiny performs best task, achieving highest accuracy ( IoU 0.99672, Dice Coefficient 0.99836). In tracking combining...