Today, the medical field has provided us with many facilities that make our lives much better than before. There have been many improvements. In medical science, whether it is surgery, medicine, ultrasound, X-rays, etc. In surgery, to reduce pain during surgery, generally, a local anesthetic is administered to a patient to numb only the area around the operation while the patient remains conscious. This method is known as peripheral nerve blocking (PNB). Even for experts in the field, identification remains challenging because the generated ultrasound images contain echo perturbations and speckle noise that makes it non-trivial to identify nerve structures. A novel approach is proposed in this paper to address challenges in medical imaging. An optimized ResU-Net variation is introduced to enhance ultrasound nerve segmentation of the brachial plexus. This approach integrates median filtering to reduce speckle noise and employs Dense Atrous Convolution (DAC) and Residual Multi-kernel Pooling (RMP) modules to improve segmentation accuracy. Although noise remains a problem in ultrasound imaging, the methods mentioned earlier help locate nerves precisely, improving patient experience and ease of treatment.