Members: Pavel Štarha, Jana Procházková, Pavel Loučka

Our research team actively investigates and implements standard image processing methods such as filtering and segmentation, with an emphasis on optimizing performance and accuracy. In the area of filtering, we focus on classical techniques such as Gaussian filtering, median filtering, or Laplacian operation, and explore their applications in noise removal and image enhancement. Segmentation is another important area of our interest, where we utilize methods such as thresholding, morphological operations, and edge detectors for identifying and separating individual objects or structures in the image.

As well as traditional methods, our research team is also exploring the use of artificial intelligence methods for image analysis in collaboration with the AI research group. We use deep learning and neural networks to enhance the capabilities of our systems, particularly in areas such as object detection, semantic segmentation and image classification. By combining classical techniques with the power of neural networks, we aim to develop more robust and adaptive solutions that can process diverse visual data with greater accuracy and efficiency. This hybrid approach allows us to harness the power of both traditional image processing methods and modern deep learning techniques.

Our research team also specializes in the development of advanced methods for processing 3D data from various sources such as 3D scanners, lidars, and x-rays. We use algorithms based on edge detectors, complementary image segmentation, and pattern recognition to detect objects and structures in 3D data, enabling the identification and extraction of key features from spatial data. These methods allow us to analyze and visualize 3D scenes with high accuracy, which has wide applications in various fields from industry to scientific research.