Martin Mirbauer
Supervisor(s): doc. Ing. Jaroslav Křivánek Ph.D.
Charles University
Abstract: Digital photography has become widespread with the global use of smartphones. However, most of the captured images do not fully use the camera capabilities and only store the captured photos in a format with limited dynamic range. Recently several new dynamic range reconstruction methods using convolutional neural networks (CNNs) were published and their performance has not yet been comprehensively compared. In this paper we present a framework for comparing dynamic range reconstruction approaches and use it to compare the reconstruction quality of the recently published CNN-based approaches. We also implement a mobile HDR camera application and evaluate the feasibility of running the best-performing reconstruction method directly on a mobile device.
Keywords: Image Processing
Full text: Year: 2019
