• Tanmay Shrivastava
• Piyush Chaube
• Aradhya Pise
• Mihir Salvi

A useful low-light enhancement method must be memory- and computationally-efficient while producing a pleasing restoration. Concerns concerning their applicability in the actual world are raised by the fact that most available approaches prioritise restoration quality while sacrificing speed and memory needs. For extreme low-light single picture restoration, we suggest a new deep learning architecture that, despite its quick and light inference, achieves a restoration that is on par perceptually with the most advanced computationally intensive models. When possible, we avoid the intermediate scales in favour of processing most of the data in the larger scale areas. Being CNN- grounded styles generally operate either on full- resolution or on precipitously low- resolution representations. In the former case, spatially precise but contextually less robust results are achieved, while in the ultimate case, semantically dependable but spatially less robust results are generated. In this paper, we present a new armature with the collaborative pretensions of maintaining spatially precise high- resolution representations through the entire network, and entering strong contextual information from the low- resolution representations. The core of our approach is a multi-scale residual block containing several crucial rudiments(a) parallel multi-resolution complication aqueducts for rooting multi-scale features,(b) information exchange across the multi-resolution streams (c) Mechanisms for acquiring contextual information that are spatial and channel-based, and (d) Multi-scale feature aggregation based on attention.

Image Enhancement,CNN

"Image Enhancement Using Deep Learning", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.9, Issue 11, page no.f168-f172, November-2022, Available :http://www.jetir.org/papers/JETIR2211522.pdf

"Image Enhancement Using Deep Learning", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.9, Issue 11, page no. ppf168-f172, November-2022, Available at : http://www.jetir.org/papers/JETIR2211522.pdf