Published Paper Details:

Even though humans performance is excellent still humans can't reach machines mechanism in visualizing the objects and images. Most of the new rediscovery of Neural Networks which has led to a notable performance and advancements in automated visual classification and their generalization capabilities are not up to human expectations, since machines learn biased feature space, which majorly turn on the working of the training datasets despite of its general principles. This project mainly inspects about the define classification of image-based processing of electroencephalogram [EEG] signals caused by the observer's brain. In our project the main chore was to focus on the particular task of processing the EEG signals resulted in brain when an image is spotted, so the kind of object appeared in the image will be well defined. Proposed methodology focuses at interpreting an input multichannel temporal EEG sequence into a low proportions feature vector summing up the relevant content for an input sequence by directly connecting the time sequences from innumerable channels into a unique feature vector, neglecting temporal.