Robust approach to detection of bubbles based on images analysis

Bubbles detection is important in various applications in areas including medicine, process control, geochemistry. The application of computer vision methods enables robust bubbles detection and classification even in complex image registration environments. Complex image background is one of the key issues we studied. Proposed method uses image segmentation based on graph cuts algorithm. Haar wavelet transform algorithm is applied on feature selection stage. The efficiency has been optimized for a continuous update of a list of voting points based on the accumulator size and position of bubbles. The efficency of proposed approach is demonstrated on the real dataset.