Technical Briefs

Imaging and Video

Thursday, 29 November 14:15 - 16:00 |  Garnet 219

Edge-Preserving Image Decomposition using L1 Fidelity with L0 Gradient - Picture

Edge-Preserving Image Decomposition using L1 Fidelity with L0 Gradient

We present an image decomposition method using L1 fidelity term
with L0 norm of gradient to decompose an image into base layer
and detail layer. Our image decomposition method can
be regarded as the fundamental tool to generate multiple image edit-
ing applications.

Chih-Tsung Shen, National Taiwan University
Feng-Ju Chang, CITI, Academia Sinica
Yi-Ping Hung, National Taiwan University
Soo-Chang Pei, National Taiwan University


Video Enhanced Gigapixel Panoramas - Picture

Video Enhanced Gigapixel Panoramas

We present a framework for embedding videos within gigapixel imagery. The combination of high-resolution imagery and video enables users to pan and zoom across the gigapixel panorama. The sparsity of the video content within the gigapixel context introduces artifacts which we overcome by optimizing the traversal of the scene.

Soeren Pirk, University of Konstanz
Michael Cohen, Microsoft Research Redmond
Oliver Deussen, University of Konstanz
Matt Uyttendaele, Microsoft Research Redmond
Johannes Kopf, Microsoft Research Redmond


Gradient Domain Salience-preserving Color-to-gray Conversion - Picture

Gradient Domain Salience-preserving Color-to-gray Conversion

This paper presents a new gradient domain color-to-gray conversion algorithm. By enhancing the luminance gradient with the chromatic difference in CIELAB space, artifact-free and salience-preserving grayscale images can be generated. The algorithm also shows perfect color discriminability and ordering for isoluminance colors.

Bingfeng Zhou, Institute of Computer Science and Technology, Peking University
Jie Feng, Institute of Computer Science and Technology, Peking University


Optimized Anaglyph Colorization - Picture

Optimized Anaglyph Colorization

An optimization framework for anaglyph recoloring, bring in new experience for viewers.

Hong-Shiang Lin, Taiwan University
Chih Lin Zheng, National Taiwan University
Yu Hsun Lin, National Taiwan University
Ming Ouhyoung, National Taiwan University


GPU-based local-dimming for power efficient imaging - Picture

GPU-based local-dimming for power efficient imaging

A physically based local-dimming algorithm has been implemented on the GPU of a notebook. No additional HW is required. A power saving rate of 34% could be achieved. This algorithm may increase the runtime of a notebook/tablet and allow the battery to be smaller, lighter and cheaper.

Chihao Xu, Saarland University
Michael Krause, Saarland University
Jens Krüger, IVDA, Intel VCI