Virtual DSLR: High Quality Dynamic Depth-of-Field Synthesis on Mobile Platforms

Shallow depth-of-field (DoF) and a smooth bokeh are signature elements of Digital SLR cameras and high-quality lenses. Producing comparable effects on mobile platforms has long been challenging due to small sensor sizes and short focal lengths of mobile cameras. In this paper, we exploit depth sensing capabilities on emerging mobile devices and develop a new depth-guided refocus synthesis technique particularly tailored for mobile devices. Our technique takes coarse depth maps as inputs and applies novel depth-aware pseudo ray tracing. The depth maps can be obtained frommobile depth sensors, mobile stereo cameras and even from user-inputs. Our pseudo ray tracing scheme resembles light field synthesis and refocusing but does not require actual creation of the light field and hence reduces both memory and computational overhead. At the same time, the scheme can overcome boundary bleeding and discontinuity artifacts observed in previous filtering techniques. Comprehensive experiments show that our approach can produce very high quality DoF comparable to the ones produced by DSLR and the Lytro light field cameras.. Document Recognition and Retrieval XXIII Intelligent Pen: A least cost search approach to stroke extraction in historical documents Kevin L. Bauer and William Barrett, Brigham Young University (USA) Abstract: Extracting strokes from handwriting in historical documents provides high-level features for the challenging problem of handwriting recognition. Such handwriting often contains noise, faint or incomplete strokes, strokes with gaps, overlapping ascenders and descenders and competing lines when embedded in a table or form, making it unsuitable for local line following algorithms or associated binarization schemes. We introduce Intelligent Pen for piece-wise optimal stroke extraction. Extracted strokes are stitched together to provide a complete trace of the handwriting. Intelligent Pen formulates stroke extraction as a set of piece-wise optimal paths, extracted and assembled in cost order. As such, Intelligent Pen is robust to noise, gaps, faint handwriting and even competing lines and strokes. Intelligent Pen traces compare closely with the shape as well as the order in which the handwriting was written. A quantitative comparison with an ICDAR hand-written stroke data set shows Intelligent Pen traces to be within 2.58 pixels (mean difference) of the manually created strokes. papers continue on page 8 Extracting strokes from handwriting in historical documents provides high-level features for the challenging problem of handwriting recognition. Such handwriting often contains noise, faint or incomplete strokes, strokes with gaps, overlapping ascenders and descenders and competing lines when embedded in a table or form, making it unsuitable for local line following algorithms or associated binarization schemes. We introduce Intelligent Pen for piece-wise optimal stroke extraction. Extracted strokes are stitched together to provide a complete trace of the handwriting. Intelligent Pen formulates stroke extraction as a set of piece-wise optimal paths, extracted and assembled in cost order. As such, Intelligent Pen is robust to noise, gaps, faint handwriting and even competing lines and strokes. Intelligent Pen traces compare closely with the shape as well as the order in which the handwriting was written. A quantitative comparison with an ICDAR hand-written stroke data set shows Intelligent Pen traces to be within 2.58 pixels (mean difference) of the manually created strokes. papers continue on page 8 All Proceedings Papers from EI 2016 are available for FREE download at www.ingentaconnect.com/content/ist/ei * These papers were presented within the conference noted at the IS&T International Symposium on Electronic Imaging, held Feb. 14–18, 2016, in San Francisco, CA.