Images with great quality are necessary for great vision, remote detecting, and medical purposes. This project proposes a method to give a dynamic range of enhancement and power constrain for images and videos. Firstly the captured image largely varies with the human viewing, therefore improvements are very essential, which are performed by multi-scale retinex (MSR) methods that exists in the literature. Sub-band and power controlled logarithm function is introduced in the MSR to include the power conservation when displayed through light emitting diode (LED). We exploit the human eye phenomenon that is human eye cannot perceive enhancement if contrast is enhanced in the darker regions of an image. We adjust contrast levels based upon the decided threshold value for image to achieve the power efficiency. The same method can also be implemented for video sequences and efficiency can be achieved. We show, using the proposed method, that we can achieve better enhancement with lower power consumption.