Introduction ➢This research explores new applicabilities of Brain Computer Interfaces (BCI) within the multimedia field. Particularly, we present system to generate a new way of saliency maps purely based on user’s brain reaction after seen a patch of an image. EEG processing and P300 detector ➢EEG signals related to each window presentation (from 200ms to 900ms after the window presentation) were extracted, band-pass filtered and down sampled to generate a feature vector. Object detection based on EEG-saliency maps ➢Maps binarized are combined with Grabcut to generate an accurate object segmentation. ➢Averaging maps across different reduce noise on the maps