Steffen Wachenfeld

Robust recognition of 1-D barcodes using camera phones

In this paper we present an algorithm for the recognition of 1D barcodes using camera phones, which is highly robust regarding the the typical image distortions. We have created a database of barcode images, which covers typical distortions, such as inhomogeneous illumination, reflections, or blurriness due to camera movement. We present results from experiments with over 1,000 images from this database using a Matlab implementation of our algorithm, as well as experiments on the go, where a Symbian C++ implementation running on a camera phone is used to recognize barcodes in daily life situations. The proposed algorithm shows a close to 100% accuracy in real life situations and yields a very good resolution dependent performance on our database, ranging from 90.5% (640 × 480) up to 99.2% (2592 × 1944). The database is freely available for other researchers.

Recognition of Screen-Rendered Text

The recognition of screen-rendered text is to our knowledge a yet unaddressed task. It has to be performed e.g. by translation tools which allow users to click on any text on the screen and give a translation. This often requires to capture a screenshot and to perform optical character recognition which is very challenging due to very small and smoothed fonts. This paper presents a method capable of recognizing smoothed and non-smoothed screen-rendered text of very small size which also works for colored fonts on inhomogeneous backgrounds

Robust 1-d barcode recognition on camera phones and mobile product information display

In this paper we present a robust algorithm for the recognition of 1-D barcodes using camera phones. The recognition algorithm is highly robust regarding the typical image distortions and was tested on a database of barcode images, which covers typical distortions, such as inhomogeneous illumination, reflections, or blurs due to camera movement. We present results from experiments with over 1,000 images from this database using a MATLAB implementation of our algorithm, as well as experiments on the go, where a Symbian C++ implementation running on a camera phone is used to recognize barcodes in daily life situations. The proposed algorithm shows a close to 100% accuracy in real life situations and yields a very good resolution dependent performance on our database, ranging from 90.5% (640 ×480) up to 99.2% (2592 ×1944). The database is freely available for other researchers. Further we shortly present MobilePID, an application for mobile product information display on web-enabled camera phones. MobilePID uses product information services on the internet or locally stored on-device data.

Graph-based markerless registration of city maps using geometric hashing

Recently, augmenting paper maps with additional dynamic information on mobile devices has become popular. A central task in this context is to register high-resolution paper maps to digital maps on a mobile device, which was typically performed by means of RFID tags or visual markers on specially prepared paper maps. In this paper we present a novel graph-based approach for a markerless registration of city maps. The goal is to find the best registration between a given image, which shows a small part of a city map, and stored map data. The proposed method creates a graph representation of a given input image and robustly finds an optimal registration using a geometric hashing technique. It is translation, scale and rotation invariant, and robust against noise and missing data. Experiments on both synthetic and real data are presented to demonstrate the algorithmic performance.

Annotated Databases for the Recognition of Screen-Rendered Text

The recognition of screen-rendered text is a novel task. It is performed e.g. by translation tools which allow users to click on any text on the screen and give a translation. Also some commercial OCR programs start to address the problem of reading screenshots. Optical character recognition on screen-shot images can be very challenging due to very small and smoothed fonts. In order to build and compare recognition approaches for screen-rendered text, the availability of standard databases is a fundamental prerequisite. In this paper two freely available databases are presented, one that consists of annotated screenshot images of 28080 single characters and another holding 400 words extracted from documents plus 2 400 generated isolated words. Both databases include meta-information such as x-height, font type, style and rendering conditions. At the example of a developed recognition system, it is shown how these databases can serve for training, testing and optimization.

Segmentation of Very Low Resolution Screen-Rendered Text

The lower the resolution of a given text is, the more difficult it becomes to segment it into single characters. The resolution of screen-rendered text can be very low. This paper focuses on smoothed screen-rendered text of very low resolution with typical x-heights of 4 to 7 pixels which is much lower than in other low resolution OCR situations. We propose a recognition-based segmentation algorithm which makes use of over segmentation by dynamic programming, candidate rating by single character classifiers and a graph based search algorithm for an optimal cut sequence. The algorithm is described in detail and experimental results are presented which show the performance on example screen- shot images taken from the public Screen-Word database.

Developing mobile multimedia applications on symbian OS devices

For the implementation of multimedia applications on mobile devices several platforms exist. Getting started with mobile programming can be difficult and tricky. This paper gives an introduction to the development of mobile multimedia applications on Symbian OS devices. In particular this paper presents the step-by-step development of an easy image processing application that makes use of the mobile devices camera. Here, a Nokia N95 phone is used, which is widely used in the community due to its features such as autofocus, GPS, wireless LAN, gravity sensors and more. This paper will provide background knowledge as well as step-by-step explanations of all implementation steps ranging from installation of an appropriate SDK to signing, installing, and running the developed application on a mobile device.

A multiple classifier approach for the recognition of screen-rendered text

The lower the resolution of a given text is, the more difficult it becomes to segment and to recognize it. The resolution of screen-rendered text can be very low. With a typical x-height of 4 to 7 pixels it is much lower as in other low resolution OCR situations. Modern OCR approaches for such very low resolution text use a classification-based segmentation where the underlying classifier plays an important role. This paper presents a multiple classifier system for the classification of single characters. This system is used as a subsystem for the classification-based segmentation within a system to read screen-rendered text. The paper shows that the presented multiple classifier system outperforms the best former single classifier system on single characters by far and it shows the impact of using the multiple classifier system on the word reading performance.

Graph-Based Registration of Partial Images of City Maps Using Geometric Hashing

In this paper, we present a novel graph-based approach for the registration of city maps. The goal is to find the best registration between a given image, which shows a small part of a city map, and stored map data. Such registration is important in fields like mobile computing for augmentation purposes. Until now, RFID tags, markers, or regular dot grids on specially prepared maps are typically required. In this paper we propose a graph-based method to avoid the need of special maps. It creates a graph representation of a given input image and robustly finds an optimal registration using a geometric hashing technique. Our approach is translation, scale and rotation invariant, map type independent and robust against noise and missing data.

DocXS distributed operator construction and execution system

In the field of computer vision and pattern recognition, data processing and data analysis tasks are often implemented as a consecutive or parallel application of more-or-less complex operations. In the following we will present DocXS, a computing environment for the design and the distributed and parallel execution of such tasks. Algorithms can be programmed using an Eclipse-based user interface, and the resulting Matlab and Java operators can be visually connected to graphs representing complex data processing workflows. DocXS is platform independent due to its implementation in Java, is freely available for noncommercial research, and can be installed on standard office computers. One advantage of DocXS is that it automatically takes care about the task execution and does not require its users to care about code distribution or parallelization. Experiments with DocXS show that it scales very well with only a small overhead.