For cinematic and episodic productions, on-set look management is an important component of the creative process, and
involves iterative adjustments of the set, actors, lighting and camera configuration. Instead of using the professional
motion capture device to establish a particular look, the use of a smaller form factor DSLR is considered for this purpose
due to its increased agility. Because the spectral response characteristics will be different between the two camera
systems, a camera emulation transform is needed to approximate the behavior of the destination camera. Recently, twodimensional
transforms have been shown to provide high-accuracy conversion of raw camera signals to a defined
colorimetric state. In this study, the same formalism is used for camera emulation, whereby a Canon 5D Mark III DSLR
is used to approximate the behavior a Red Epic cinematic camera. The spectral response characteristics for both cameras
were measured and used to build 2D as well as 3x3 matrix emulation transforms. When tested on multispectral image
databases, the 2D emulation transforms outperform their matrix counterparts, particularly for images containing highly
The dead leaves model was recently introduced as a method for measuring the spatial frequency response (SFR) of
camera systems. The target consists of a series of overlapping opaque circles with a uniform gray level distribution and
radii distributed as r-3. Unlike the traditional knife-edge target, the SFR derived from the dead leaves target will be
penalized for systems that employ aggressive noise reduction. Initial studies have shown that the dead leaves SFR
correlates well with sharpness/texture blur preference, and thus the target can potentially be used as a surrogate for more
expensive subjective image quality evaluations. In this paper, the dead leaves target is analyzed for measurement of
camera system spatial frequency response. It was determined that the power spectral density (PSD) of the ideal dead
leaves target does not exhibit simple power law dependence, and scale invariance is only loosely obeyed. An extension
to the ideal dead leaves PSD model is proposed, including a correction term to account for system noise. With this
extended model, the SFR of several camera systems with a variety of formats was measured, ranging from 3 to 10 megapixels;
the effects of handshake motion blur are also analyzed via the dead leaves target.
We propose a system for display of large format images such that natural scenes can be approximated in a laboratory
setting, both spectrally and in dynamic range. The system uses a large area (48"x36") high intensity light box
constructed using multiple xenon arc lamps with diffusers to maximize surface uniformity. This source is used as a back
light for multiple transparency layers mounted on a rigid plexiglass substrate, where images are printed on the
transparency layers using a wide format inkjet printer. A detailed transparency printer model for this system is described,
where the spectral transmittance can be predicted with a high degree of accuracy for any given combination of input
digital ink values (CMYKRGB). This spectral printer model, in conjunction with knowledge of the spectral
characteristics of the back light, can be used to approximate the spectral radiance of a target scene. On a per-pixel level,
nonlinear constrained optimization is used to solve for the combination of printer inks that produces the best estimate to
the spectrum of the target scene. With this system, it is then possible to create realistic static images with a large
dynamic range that can be used to benchmark camera systems in a controlled laboratory setting.
Inherent to most multi-color printing systems is the inability to achieve perfect registration between the primary
separations. Because of this, dot-on-dot or dot-off-dot halftone screen sets are generally not used, due to the
significant color shift observed in the presence of even the slightest misregistration. Much previous work has
focused on characterizing these effects, and it is well known that dot-off-dot printed patterns result in a higher
chroma (C*) relative to dot-on-dot. Rotated dot sets are used instead for these systems, as they exhibit a much
greater robustness against misregistration. In this paper, we make the crucial observation that while previous
work has used color shifts caused by misregistration to design robust screens, we can infact exploit these color
shifts to obtain estimates of misregistration. In particular, we go on to demonstrate that even low resolution
macroscopic color measurements of a carefully designed test patch can yield misregistration estimates that are
accurate up-to the sub-pixel level. The contributions of our work are as follows: 1.) a simple methodology to
construct test patches that may be measured to obtain misregistration estimates, 2.) derivation of a reflectance
printer model for the test patch so that color deviations in the spectral or reflectance space can be mapped to
misregistration estimates, and 3.) a practical method to estimate misregistration via scanner RGB measurements.
Experimental results show that our method achieves accuracy comparable to the state-of-the art but expensive
geometric methods that are currently used by high-end color printing devices to estimate misregistration.
A new method for halftoning using high resolution pattern templates is described, that expands the low level rendering capabilities for engines that support this feature. This approach, denoted super resolution encoded halftoning (SREH) is an extension of the Holladay concept, and provides a compact way to specify high resolution dot growth patterns using a lower resolution Holladay brick. Fundamentally, this new halftoning method involves using the SRE patterns as building blocks for constructing clustered dot growth assemblies. Like the traditional Holladay dot description, the SRE halftone is characterized by a size, height, and shift, all of which are specified at the lower resolution. Each low resolution pixel position in the SRE halftone brick contains a pair of lists. The first of these is a list of digital thresholds at which a transition in SRE patterns occurs for that pixel position, and the second is the corresponding list of SRE codes. For normal cluster dot growth sequences, this provides a simple and compact mechanism for specifying higher resolution halftones. Techniques for emulating traditional high resolution Holladay dots using SREH are discussed, including mechanisms for choosing substitutions for patterns that do not exist among the available SRE patterns.
We present a new approach for achieving spatial frequency filtering in the analog domain. Our device, the Thin Film Analog Image Processor (TAIP), is a hybrid structure that combines the strengths of analog VLSI technology with the simplicity of a conducting polymer film. The TAIP consists of a silicon chip with a square array of metal pads corresponding to the image pixels, onto which a conducting polymer film is applied to create lateral interaction between pixels. Analog image data (0 - 2 Volts) is multiplexed into the array, and the image is processed with up to 72 dB of resolution. The TAIP arrays are capable of performing either high or low pass spatial frequency filtering, operations that become computationally intensive for large images in the digital domain. Multiple arrays can be combined to create tunable bandpass spatial frequency filters that are capable of extracting features from complex images. Two array formats have been fabricated, 60 X 80 and 320 X 240, each capable of 60 Hz operation with a power consumption of approximately 100 mW.