A Hybrid GRU-CNN Feature Extraction Technique for Speaker Identification