1. |
- Assabie, Yaregal, 1975-, et al.
(författare)
-
A comprehensive Dataset for Ethiopic Handwriting Recognition
- 2009
-
Ingår i: Proceedings SSBA '09. - Halmstad : Halmstad University. - 9789163339240 ; , s. 41-43
-
Bokkapitel (övrigt vetenskapligt/konstnärligt)abstract
- Ethiopic script is used by several languages in Ethiopia for writing. We present a comprehensive dataset of handwritten Ethiopic script called DEHR (Dataset for Ethiopic Handwriting Recognition) captured both offline and online. The offline dataset includes isolated characters, Ethiopian church documents and ordinary handwritten texts dealing with various real-life issues. The ordinary texts and isolated characters were freely written by several participants. The church documents are written in Geez and Amharic languages whereas the language for ordinary texts is Amharic only. The online dataset was collected by using two Digimemo devices of different sizes. For isolated characters and online dataset, all the 265 character samples used by Amharic language are included. The dataset is intended to set a benchmark for training and/or testing handwriting recognition, character and word segmentation, and text line detection. The dataset is can be accessed by contacting the authors or via http://www.hh.se/staff/josef/.
|
|
2. |
- Assabie, Yaregal, 1975-, et al.
(författare)
-
Offline Handwritten Amharic Word Recognition Using HMMs
- 2009
-
Ingår i: Proceedings SSBA '09. - Halmstad : Halmstad University. - 9789163339240 ; , s. 89-92
-
Bokkapitel (övrigt vetenskapligt/konstnärligt)abstract
- This paper describes two appraches for Amharic word recognition in unconstrained handwritten text using HMMs. The first approach builds word models from concatenated features of constituent characters and in the second method HMMs of constituent characters are concatenated to form word model. In both cases, the features used for training and recognition are primitive strokes and their spatial relationships. The recognition system does not require segmentation of characters but requires text line detection and extraction of structural features, which is done by making use of direction field tensor. The performance of the recognition system is tested by DEHR dataset of unconstrained handwritten documents collected from various sources.
|
|
3. |
- Gavrilovic, Milan, 1981-, et al.
(författare)
-
Dimensionality Reduction for Colour Based Pixel Classification
- 2009
-
Ingår i: Proceedings SSBA 2009. - Halmstad : Halmstad University. - 9789163339240 ; , s. 65-68
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- In digital images, providing classification based on colour, hue or spectral angle is a problem usually solved by combining a variety of pre-processing steps, as well as object wise classifiers. We have developed a method for transforming colour or multispectral image data to a 1D colour histogram with respect to the digital characteristics of intensity measurements. Classification is then reduced to 1D histogram segmentation which is a simpler problem. The proposed method, based on ideas of spectral decomposition, was previously applied in dual-colour fluorescence microscopy for quantification and detection of colocalization insensitive to cross-talk. In this paper the principle is expanded to unsupervised colour based pixel classification algorithms in hue-saturation-lightness or luminance-chrominance colour spaces.
|
|
4. |
- Kolbe, Viktor, et al.
(författare)
-
Indoor Photorealistic 3D Mapping using Stereo Images from SLR Cameras
- 2009
-
Ingår i: Proceedings of the '09 Swedish Symposium on Image Analysis (SSBA). - Linköping : Linköping University Electronic Press. - 9789163339240
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- Creating a 3D model from photos require an estimate of the position and orientation (pose) of the camera for each photo that is acquired. This paper presents a method to estimate the camera pose using only image data. The images are acquired at a low frequency using a stereo rig, consisting of two rigidly attached SLR cameras. Features are extracted and an optimization problem is solved for each new stereo image. The results are used to merge multiple stereo images and building a larger model of the scene. The accumulated error after processing 10 images can with the present methods be less than 1.2 mm in translation and 0.1 degrees in rotation.
|
|
5. |
- Kylberg, Gustaf, et al.
(författare)
-
Detecting Virus-like Particles Using Transmission Electron Microscopy
- 2009
-
Ingår i: Proceedings SSBA 2009, Symposium on Image Analysis. - Halmstad : EIS, Halmstad University. - 9789163339240 ; , s. 13-16
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- We present a multi scale approach for automating the image acquisition process for an computerized virus diagnostic application. Our methods are designed to mimic the methodology used by virus TEM experts manually operating the microscope. The methods decrease the search area considerably. In addition we present a segmentation method for virus-like particles based on local intensity information and PCA. This method makes no assumption regarding shape which is vital since many viruses are highly pleomorphic, i.e., have different shapes. The only input parameter used is the approximate virus thickness, which is a conserved feature within a virus species.
|
|
6. |
|
|
7. |
- Malm, Patrik, 1982-, et al.
(författare)
-
Curve closing by gradient weighted distance transformation applied in automated cervical cancer screening
- 2009
-
Ingår i: Proceedings SSBA 2009. - 9789163339240 ; , s. 25-28
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- In this paper, a nuclei segmentation algorithm that uses a gradient weighted distance transform for curve closing is described. Edge detection methods often return broken edges that need to be closed to achieve proper segmentation. Based n a Canny edge result, this method generates a distance map that is dependent on local gradient directions and magnitudes making it possible to promote propagation along edges rather than away from them. Through this procedure, which is a kind of adaptive morphology, it is possible to achieve curve closing even along very weak edges. The method is developed within a larger project aimed at creating a cervical cancer screening system mainly intended to allow automated screening in developing countries.
|
|
8. |
- Malmberg, Filip, et al.
(författare)
-
Interactive Segmentation with Relaxed Image Foresting Transforms
- 2009
-
Ingår i: Proceedings of SSBA 2009. - 9789163339240
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- The Image Foresting Transform (IFT) is a framework for efficient image partitioning, used in interactive segmentation. We propose a modified version of the IFT, and demonstratethat the modified algorithm is more robust to noise while maintaining computational complexity. We also show an application of the method for interactive segmentation ofback muscles in magnetic resonance images, where seedpoints representing object and background are placed repeatedly until a desired segmentation result is obtained.
|
|
9. |
- Selig, Bettina, 1982-, et al.
(författare)
-
Measuring Distribution of Lignin in Wood Fibre Cross-Sections
- 2009
-
Ingår i: Proceedings SSBA 2009. - Halmstad : EIS, Halmstad University. - 9789163339240 ; , s. 5-8
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- Lignification of wood fibres has important consequences to the paper production, but its exact effects are not well understood. To correlate exact levels of lignin in wood fibres to their mechanical properties, lignin autofluorescence is imaged in wood fibre cross-sections. Highly lignified areas can be detected and related to the area of the whole cell wall. Presently these measurements are performed manually, which is tedious and expensive. In this paper a method is proposed to estimate the degree of lignification automatically. The method is evaluated manually by an expert. Beside some difficulties segmenting cells that do not conform to our model, there was a highly significant correlation between the two methods.
|
|
10. |
- Teferi, Dereje, 1969-, et al.
(författare)
-
Evaluation Protocol for the DXM2VTS Database and Performance Comparison of Face Detection and Face Tracking on Video
- 2009
-
Ingår i: Proceedings SSBA '09. - Halmstad : Halmstad University. - 9789163339240 ; , s. 49-52
-
Bokkapitel (övrigt vetenskapligt/konstnärligt)abstract
- Video databases and their corresponding evaluation protocols are used to compare classifiers, such as face detection an tracking. In this paper, a six level evaluation protocol for the Damascened XM2VTS (DXM2VTS) database is presented to measure face detection and tracking performance. Additionally, a novel database containing thousands of videos is created by combining video from the XM2VTS database with a set of newly recorded standardized real-life video used as background and with several realistic degradations, such as motion blur, noise, etc. Moreover, two publicly available and published face detection algorithms, are tested on the six suggested difficulty levels of the protocol. Their performance on video in terms of false acceptance, false rejection, correct detection, and repeatability, are reported and conclusions are drawn.
|
|