Publications Repository - Gdańsk University of Technology

in the paper, after pointing out of realistic recordings and classifications of their frames, we propose a new shifting time window approach for improving binary classifications. We consider image classification in tewo steps. in the first one the well known binary classification algorithms are used for each image separately. In the second step the results of the previous step mare analysed in relatively short sequences of consecutive images. Taking into account the nature of the recorded and analysed video frame sequences, some corrections of the binary classification sequences can be done. In consequence the quality of the considered classification algorithm can be improved. Furthermore, we offer a systematic approach where all parameters of the proposed algorithm such as the window lenght or parmeters of the vote in the window can be considered and their optimal values properly determined. we also present representative examples which confirm the advantages of the proposed approach.

Authors

• mgr inż. Adam Blokus link open in new tab ,
• prof. dr hab. inż. Henryk Krawczyk link open in new tab