Agata Chmielewska

Influence of low resolution of images on reliability of face detection and recognition

In this paper we analyze reliability of the real-time system for face detection and recognition from low-resolution images, e.g., from video monitoring images. First, we briefly describe main features of the standards for biometric face images. Available scientific databases have been checked for compliance with these biometric standards. During the research we have considered both the correctness of extraction (location) of the face from the image as well as the correctness of the identification (based on the eigenfaces approach). To the tests we have used the face databases that allow to study tolerance to illumination and face positions. We have compared various face detection techniques and analyzed minimum requirements for the resolution of facial images. Finally, an influence of the resolution reduction on the FAR/FRR of the recognition is presented.

Selection of parameters in iris recognition system

This paper presents the detailed analysis of implementation issues occurred during preparation of the novel iris recognition system. First, we shortly describe the currently available acquisition systems and databases of iris images, which were used for our tests. Next, we concentrate on the feature extraction and coding with the execution time analysis. Results of the average execution time of loading the image, segmentation, normalization, and feature encoding, are presented. Finally, DET plots illustrate the recognition accuracy for IrisBath database.

Face Recognition from Low Resolution Images

This paper describes an analysis of the real-time system for face recognition from video monitoring images. First, we briefly describe main features of the standards for biometric face images. Available scientific databases have been checked for compliance with these biometric standards. Next, we concentrate on the analysis of the prepared face recognition application based on the eigenface approach. Finally, results of our face recognition experiments with images of reduced resolution are presented. It turned out that the proposed and tested algorithm is quite resistant to changing the resolution. The recognition results are acceptable even for low-resolution images (16×20 pixels).

New approach to traffic density estimation based on indoor and outdoor scenes from CCTV

Abstract In this paper we present an algorithm for precise estimation of moving objects density (typically people and vehicles) in indoor and outdoor scenes. Automatic generation of the so-called density maps is based on video sequences acquired by surveillance systems. Our approach offers two types of solutions. The first one increments the accumulation table when a moving object is detected in a location of interest, delivering a density map of the presence of moving objects. The second algorithm increments the accumulation table only in cases of detecting a new moving object, resulting in a density map of the count of moving objects. The proposed algorithms were tested with the use of PETS 2009 database and with our own database of long-term video recordings. Finally, results of the density maps visualization and determination of the “busy hours” are presented.

Application of the projective geometry in the density mapping based on CCTV monitoring

In this paper we address a problem of automatic generation of the so-called density maps (heat maps) using video recordings acquired from surveillance CCTV systems in urban areas (CCTV - closed circuit television). Our software for density map estimation increments the accumulation table only in case of detecting a new moving object, resulting in a count of moving objects density map. We also applied an algorithm for projective transformation of image to achieve better accuracy of the moving object density maps. We analyzed the proposed solution for different angles of camera positions. Next the results were compared with the density maps calculated from the top view camera position.

Approach to local contrast enhancement

In this article a simple and fully automatic approach to local contrast enhancement is presented, maintaining reasonable computational complexity and feasibility of simultaneous calculations on multiple processors. The authors propose an algorithm, which allows to obtain images with artificially increased dynamic range. The resulting images do not contain unnatural artifacts and are close to images perceived by humans. Results of experiments show that the described method offers very good results even for images obtained with a large range of differences in the exposures.

Analysis of Particular Iris Recognition Stages

In this paper particular stages are analyzed present in the iris recognition process. First, we shortly describe available acquisition systems and databases of iris images, which can be used for tests. Next, we concentrate on features extraction and coding with the time analysis. Results of average time of loading the image, segmentation, normalization, features encoding, and also recognition accuracy for CASIA and IrisBath databases are presented.