CONTRAST ENHANCEMENT SCHEME FOR VISUAL SALIENCY DETECTION

Visual saliency detection becomes more popular research topic in the recent years because the prediction of attention finds application in object detection, object recognition and human computer interface etc. This paper presents the bottom up visual saliency detection with the help of contrast enhancement. Initially the SLIC segmentation is performed as preprocessing step. The global contrast enhancement scheme is adopted in this proposed method. The color distribution and contrast enhancement map is multiplied together to get the final saliency map. The performance metrics of Precision, Recall, f-measure, ROC, AUC are obtained to validate the results. The benchmark dataset of MSRA and ECSSD are used to evaluate the performance metric of the proposed method and state-of-methods. The experimental results demonstrate that the proposed method is computationally efficient.