Multi-scale fusion for underwater image enhancement using multi-layer perceptron

<span id="docs-internal-guid-54b35aa6-7fff-0992-ed4c-aca4d05cfcfa"><span>Underwater image enhancement (UIE) is an imperative computer vision activity with many applications and different strategies proposed in recent years. Underwater images are firmly low quality by a mixture of noise, wavelength dependency, light attenuation. This paper depicts effective strategy to improve the degraded underwater images. Existing methods for dehazing literature considering dark channel prior utilize two separate phases evaluating transmission map (i.e., estimation refinement). Accurate restoration not possible these takes more computational time. A three-step method imaging approach that does need particular hardware or conditions. First, we multi-layer perceptron (MLP) comprehensively evaluate maps base channel, followed contrast enhancement. Furthermore, gamma-adjusted version MLP recovered derived. Finally, multi-scale fusion was applied attained The standardized weight computed three weights process. quantitative results show significantly our gives better result difference 0.536, 2.185, 1.272 PCQI, UCIQE, UIQM metrics, respectively, on single benchmark dataset. qualitative also give compared state-of-the-art techniques.</span></span>