A novel hybrid technique for low exposure image enhancement using sub-imge histogram equilization and artificial neural network
Manveer Kaur, Karun Verma · 2016
Capturing of the images through smart phones or digital cameras has become very common activity in recent years. Even if there had been a lot of advancement in the cameras' quality in the recent years, but still the images taken by them in underwater medium or dim light conditions are subject to low exposure problem. This conditions, results in low contrast images where the finer and the valuable information is often lost for further processing. Hence, their enhancement is necessary to retrieve the information of the interest. The proposed method is hybrid method that is very effective in enhancing the images taken in underwater medium and in dim light conditions such as night vision images. The method undergoes three main steps: first, the image that is to be enhanced is passed to neural network which will give the value by which the enhancement in the image can be done. In second step, the clipping of the histogram is done according to the value calculated in the first step. In third step: the image's histogram is divided into four parts based upon exposure values then histogram equalization method is applied to all the four parts individually. The proposed method is assessed on the basis of three parameters: PSNR, MSE and Entropy.