In recent years, diagnosing neurogenerative disorders has played a vital part in the healthcare field especially Alzheimer’s Disease (AD) caused due to aging. The human brain’s limbic system includes the hippocampal region, which is crucial for memory formation and managing intellectual capacity. Alzheimer’s disease at its early stage can cause changes in the hippocampal region. Since there are no effective medications for dementia, but it is possible to take faster preventive and therapeutic action if hippocampal alterations are detected through Magnetic Resonance Imaging (MRI) during its first occurrence. The complex architecture of the hippocampus makes it impossible to segment using conventional image segmentation techniques. Among the finest solutions, automated segmentation methods based on deep learning are used to segment the hippocampus region accurately within a realistic timeframe from the brain MR images. Even though several automatic hippocampus segmentation methods have been developed in deep learning to infer the accuracy, its training time is an important factor when a huge amount of data is to be trained. Mainly, this proposed method concentrates on training time using transfer learning techniques which use the ResNet-34 as a backbone integrated with two deep learning-based semantic segmentation models namely U-Net and LinkNet architectures. The model’s performance is evaluated to learn to segment the hippocampus for 25 epochs. The training was performed in the publicly available dataset. The F1-score, IoU, Precision, and Recall of R-34-LinkNet are achieved with 0.85, 0.75, 0.884, and 0.899 respectively. For the R-34-U-Net, F1-score, IoU, Precision, and Recall are 0.85, 0.75, 0.883, and 0.858 respectively.