Fast cat-eye effect target recognition based on saliency extraction

• A fast cat-eye effect target recognition method based on saliency extraction is proposed.
• We adopt multi-core parallel computing in our method and test the running speed.
• Experiment results show the proposed method is superior to other methods in accuracy and robustness.

Background complexity is a main reason that results in false detection in cat-eye target recognition. Human vision has selective attention property which can help search the salient target from complex unknown scenes quickly and precisely. In the paper, we propose a novel cat-eye effect target recognition method named Multi-channel Saliency Processing before Fusion (MSPF). This method combines traditional cat-eye target recognition with the selective characters of visual attention. Furthermore, parallel processing enables it to achieve fast recognition. Experimental results show that the proposed method performs better in accuracy, robustness and speed compared to other methods.