The use of biometric recognition for identifying individuals has become common in modern security systems. However, the growing demand and usability of this technology have also increased the associated security risks. Therefore, it is necessary to find more reliable biometric traits than the ones currently in use. Recently, the brain signal captured using the electroencephalogram (EEG) technique has been identified as a promising biometric option due to its exceptional level of distinctiveness, consistency, and widespread application. This paper presents a novel approach to biometric identification utilizing EEG data. The methodology employs a short-time Fourier transform (STFT) for feature extraction, followed by channel-wise division of the data. Subsequently, convolutional neural network (CNN) and long short-term memory (LSTM) models are implemented individually for each channel. The final step involves the integration of these models using an ensemble voting classifier. By utilizing an ensemble method, the system achieves notable improvements in performance criteria such as accuracy, precision, recall, and f1-score compared to the baseline papers. We achieve this by combining the predictions from multiple models.