Skip to main content

Kari Pulli

Followers

59

Following

39

Public Views

Interests

Uploads

Papers by Kari Pulli

Training with the Invisibles: Obfuscating Images to Share Safely for Learning Visual Recognition Models

ArXiv, 2019

High-performance visual recognition systems generally require a large collection of labeled image... more High-performance visual recognition systems generally require a large collection of labeled images to train. The expensive data curation can be an obstacle for improving recognition performance. Sharing more data allows training for better models. But personal and private information in the data prevent such sharing. To promote sharing visual data for learning a recognition model, we propose to obfuscate the images so that humans are not able to recognize their detailed contents, while machines can still utilize them to train new models. We validate our approach by comprehensive experiments on three challenging visual recognition tasks; image classification, attribute classification, and facial landmark detection on several datasets including SVHN, CIFAR10, Pascal VOC 2012, CelebA, and MTFL. Our method successfully obfuscates the images from humans recognition, but a machine model trained with them performs within about 1% margin (up to 0.48%) of the performance of a model trained w...

The Frankencamera

Figure 1: Two implementations of the Frankencamera architecture: (a) The custom-built F2-portable... more Figure 1: Two implementations of the Frankencamera architecture: (a) The custom-built F2-portable and self-powered, best for projects requiring flexible hardware. (b) A Nokia N900 with a modified software stack-a compact commodity platform best for rapid development and deployment of applications to a large audience.

An energy efficient time-sharing pyramid pipeline for multi-resolution computer vision

2013 IFIP/IEEE 21st International Conference on Very Large Scale Integration (VLSI-SoC), 2013

We introduce an energy efficient time-sharing pyramid pipeline architecture designed for multi-re... more We introduce an energy efficient time-sharing pyramid pipeline architecture designed for multi-resolution image analysis in mobile computer vision. The time-sharing pipeline efficiently reduces the off-chip memory traffic by reorganizing the data storage and processing order of an image pyramid. We build a parameterized image pyramid hardware generator and successfully evaluate the overall pyramid design space. Our results demonstrate that the time-sharing pyramid pipeline achieves about 50% of hardware savings in terms of area and power consumption compared to the traditional linear pipeline implementation. We also implement the multi-resolution Lucas-Kanade optical flow algorithm on the time-sharing pipeline, and demonstrate an order of magnitude savings in off-chip memory traffic and system energy consumption.

Fast ANN for High-Quality Collaborative Filtering

Computer Graphics Forum, 2015

Panoramic imaging system for camera phones

2010 Digest of Technical Papers International Conference on Consumer Electronics (ICCE), 2010

We introduce a mobile system for high resolution panoramic image creation. In our system the user... more We introduce a mobile system for high resolution panoramic image creation. In our system the user can rotate the camera arbitrarily and see the preview panorama in real-time. The system automatically captures high resolution images and generates a high-quality result with unlimited viewing angle. We employ a coarse-to-fine method for high-quality registration, and a seam-finding method to remove the ghosting effect due to moving objects. The proposed system has been tested on several types of camera phones, and the tests reveal that the system can efficiently provide a high-quality panoramic image in spite of the low computational power and memory available in such devices.

Projective Surface Matching of Colored 3D Scans

Fifth International Conference on 3-D Digital Imaging and Modeling (3DIM'05)

We present a new method for registering multiple 3D scans of a colored object. Each scan is regar... more We present a new method for registering multiple 3D scans of a colored object. Each scan is regarded as a color and range image of the object recorded by a pinhole camera. Consider a pair of cameras that see overlapping parts of the objects. For correct camera poses, the actual image of the overlap area in one camera matches the rendition of the overlap area as seen by the other camera. We define a mismatch score summarizing discrepancies in color, range, and silhouette between pairs of images, and we present an algorithm to efficiently minimize this mismatch score over camera poses.

Cascaded displays

ACM Transactions on Graphics, 2014

We demonstrate that layered spatial light modulators (SLMs), subject to fixed lateral displacemen... more We demonstrate that layered spatial light modulators (SLMs), subject to fixed lateral displacements and refreshed at staggered intervals, can synthesize images with greater spatiotemporal resolution than that afforded by any single SLM used in their construction. Dubbed cascaded displays , such architectures enable superresolution flat panel displays (e.g., using thin stacks of liquid crystal displays (LCDs)) and digital projectors (e.g., relaying the image of one SLM onto another). We introduce a comprehensive optimization framework, leveraging non-negative matrix and tensor factorization, that decomposes target images and videos into multi-layered, time-multiplexed attenuation patterns---offering a flexible trade-off between apparent image brightness, spatial resolution, and refresh rate. Through this analysis, we develop a real-time dual-layer factorization method that quadruples spatial resolution and doubles refresh rate. Compared to prior superresolution displays, cascaded dis...

Techniques for generating robust stereo images

DT-SLAM: Deferred Triangulation for Robust SLAM

2014 2nd International Conference on 3D Vision, 2014

Obtaining a good baseline between different video frames is one of the key elements in vision-bas... more Obtaining a good baseline between different video frames is one of the key elements in vision-based monocular SLAM systems. However, if the video frames contain only a few 2D feature correspondences with a good baseline, or the camera only rotates without sufficient translation in the beginning, tracking and mapping becomes unstable. We introduce a real-time visual SLAM system that incrementally tracks individual 2D features, and estimates camera pose by using matched 2D features, regardless of the length of the baseline. Triangulating 2D features into 3D points is deferred until keyframes with sufficient baseline for the features are available. Our method can also deal with pure rotational motions, and fuse the two types of measurements in a bundle adjustment step. Adaptive criteria for keyframe selection are also introduced for efficient optimization and dealing with multiple maps. We demonstrate that our SLAM system improves camera pose estimates and robustness, even with purely rotational motions.

Robust image registration for multi-frame mobile applications

2010 Conference Record of the Forty Fourth Asilomar Conference on Signals, Systems and Computers, 2010

Abstract We present an approach to global image registration that allows a robust and efficient i... more

Mobile Visual Computing

2009 International Symposium on Ubiquitous Virtual Reality, 2009

Smart phones are becoming visual computing powerhouses. Using sensors such as camera, GPS, and ot... more Smart phones are becoming visual computing powerhouses. Using sensors such as camera, GPS, and others, the device can provide a new user interface to the real world, augmenting user's view of the world with additional information and controls. Combining computation with image capture allows new kind of photography that can be more expressive than is possible to obtain with a traditional camera. New APIs allow harnessing more computation power of a smart phone to visual processing than what one can obtain from just a CPU.

Motion-blur-free exposure fusion

2010 IEEE International Conference on Image Processing, 2010

We present a novel approach to HDR (high-dynamic-range) image fusion that copes with image blur d... more We present a novel approach to HDR (high-dynamic-range) image fusion that copes with image blur degradation often present in longexposed images. The proposed approach can deal with both camera and object motion blur in a computationally efficient manner suitable for implementation on a mobile device. The main idea is to exploit the differences between the image degradations that affect images captured for HDR fusion. Short-exposed images are mainly affected by sensor noise and less affected by motion blur, whereas longer exposed images are less noisy but potentially blurry due to motion during their exposure. Our approach consists of two steps. First we calculate an HDR representation of the scene by applying a typical HDR fusion approach. This representation could be blurry if some of the input images with long exposure time are blurry. We then fuse the HDR result with a photometrically modified version of the image with the shortest exposure time, which allows us to retain the sharpness of the short-exposed image and the noise-free characteristics of the HDR-fused image. The method does not assume an invariant blur PSF over any of the input images, and hence it can solve both for local and global blur, due to object or camera motion, respectively. We demonstrate the algorithm through a series of experiments and comparisons on natural images.

Multi-exposure imaging on mobile devices

Proceedings of the 18th ACM international conference on Multimedia, 2010

FCam

SIGGRAPH Asia 2011 Courses, 2011

ABSTRACT This course is designed to demonstrate the use of the FCam, an Application Programming I... more ABSTRACT This course is designed to demonstrate the use of the FCam, an Application Programming Interface (API) which allows for easy and precise control of camera systems. It lets programmers implement advanced computational photography applications. Such applications even allow the modification of basic camera routines such as metering, which are hidden inside black boxes in most camera systems. FCam is also an excellent tool for teaching and research on interactive mobile computational photography. The API has been open-sourced, and is currently implemented for Nokia N900 and an experimental Frankencamera (the F2). The Frankencamera architecture for computational cameras was presented at SIGGRAPH 2010. By end-2011, the FCam will hopefully be available on Nokia#8482; Linux phones, NVIDIA#8482; Tegra 3 development boards, and F3, the next version of the Stanford experimental camera. The F3 will be distributed to research universities within the US along with courseware in a National Science Foundation (NSF)-sponsored project. NVIDIA Tegras and Nokia#8482; Linux phones are available globally.

Panoramic imaging system for mobile devices

SIGGRAPH '09: Posters, 2009

Color matching for high-quality panoramic images on mobile phones

IEEE Transactions on Consumer Electronics, 2010

We present an efficient image stitching approach for creating high-quality continuous circular 36... more We present an efficient image stitching approach for creating high-quality continuous circular 360degree panoramic images. We address the problem of color and luminance correction for long, 360-degree image sequences, where the source images have very different colors and luminance. Color matching is used for minimizing both the color differences of neighboring images and the overall color correction over the whole sequence. We perform gamma correction for the luminance component and linear correction for the chrominance components of the source images. The color consistency problem of 360-degree panoramic images is addressed by color correspondence and color difference distribution processes. We implement the stitching approach in a panoramic imaging system for creating high-resolution and high-quality panoramic images on mobile phones 1 .

The Digital Michelangelo Project

Computer Graphics Forum, 1999

Figure 1: Renderings of the statues we scanned (except the David). Our raw database (including th... more

Metering for Exposure Stacks

Computer Graphics Forum, 2012

When creating a High‐Dynamic‐Range (HDR) image from a sequence of differently exposed Low‐Dynamic... more When creating a High‐Dynamic‐Range (HDR) image from a sequence of differently exposed Low‐Dynamic‐Range (LDR) images, the set of LDR images is usually generated by sampling the space of exposure times with a geometric progression and without explicitly accounting for the distribution of irradiance values of the scene. We argue that this choice can produce sub‐optimal results both in terms of the number of acquired pictures and the quality of the resulting HDR image.This paper presents a method to estimate the full irradiance histogram of a scene, and a strategy to select the set of exposures that need to be acquired. Our selection usually requires a smaller or equal set of LDRs, yet produces higher quality HDR images.

FCam for multiple cameras

SPIE Proceedings, 2012

The Frankencamera (FCam) architecture and API enables precise control over the camera in computat... more The Frankencamera (FCam) architecture and API enables precise control over the camera in computational photography applications. We present an extension to FCam API for systems equipped with multiple cameras. The proposed extension allows for an enumeration of cameras and their corresponding properties, such as position or orientation. In addition, we explicitly support camera synchronization, either through hardware mechanisms or software primitives. If hardware synchronization is available, cameras can be grouped together under a concept of a multi-sensor. Otherwise, multiple camera streams are scheduled asynchronously and synchronized using our software control primitives.

Real-time enveloping with rotational regression

ACM Transactions on Graphics, 2007

Enveloping, or the mapping of skeletal controls to the deformations of a surface, is key to drivi... more Enveloping, or the mapping of skeletal controls to the deformations of a surface, is key to driving realistic animated characters. Despite its widespread use, enveloping still relies on slow or inaccurate deformation methods. We propose a method that is both fast, accurate and example-based. Our technique introduces a rotational regression model that captures common skinning deformations such as muscle bulging, twisting, and challenging areas such as the shoulders. Our improved treatment of rotational quantities is made practical by model reduction that ensures real-time solution of least-squares problems, independent of the mesh size. Our method is significantly more accurate than linear blend skinning and almost as fast, suggesting its use as a replacement for linear blend skinning when examples are available.