Most of the existing document-binarization techniques deal with many parameters that require a priori setting of their values. Due to the unknown of the ground-truth images, the evaluation of document binarization tec...
详细信息
ISBN:
(纸本)9780889867192
Most of the existing document-binarization techniques deal with many parameters that require a priori setting of their values. Due to the unknown of the ground-truth images, the evaluation of document binarization techniques is subjective and employs human observers for the estimation of the appropriate parameter values. The selection of the appropriate values for these parameters is crucial and influences to the final binarization. However, there is no predetermined set of parameters that guarantees optimal binarization for all document images. This paper proposes a new technique that allows the estimation of proper parameters values for each one of the document binarization techniques. The proposed approach is based on a statistical performance analysis of a set of binarization results, which are obtained by applying various binarization techniques with different parameter values. The proposed statistical performance analysis can also depicts the best document binarization result obtained by a set of document binarization techniques.
This paper presents a ground truth data collection effort along with its use in evaluating unmixing algorithms. Unmixing algorithms are typically evaluated using synthetic data generated by selecting endmember spectru...
详细信息
ISBN:
(纸本)9780819471574
This paper presents a ground truth data collection effort along with its use in evaluating unmixing algorithms. Unmixing algorithms are typically evaluated using synthetic data generated by selecting endmember spectrums and adding them in different amounts and with added noise. Going from synthetic to real data poses many problems. One of the greatest is the amount of data to be collected. Also, there will be many unmodeled variations in real data. These include greater variation of the endmembers, additional endmembers that are a very small percentage of the image, and nonlinear effects in the data that are not modeled. The data collation effort produced a high resolution class map along with spectral measurements of 153 different sampling sites to validate the map. The methodology for using this high resolution class map for generating the ground truth data for use in the unmixing algorithms is presented. Specifically, a 1m class map is used to generate the endmember abundances for every pixel in a 30m Hyperion image of the Enrique Reef in Southwest Puerto Rico. The results using two unmixing algorithms, one with a sum to one constraint and the other with a non-negative constraint are presented. The unmixing results for each endmember are presented along with a newly developed unmixing parameter called the Correct Unmixing Index (CUI).
Audio gunshot recordings can be helpful for crime scene reconstruction, estimation of the shooter's location and orientation, and verification of eyewitness accounts. The audio evidence can include the muzzle blas...
详细信息
Audio gunshot recordings can be helpful for crime scene reconstruction, estimation of the shooter's location and orientation, and verification of eyewitness accounts. The audio evidence can include the muzzle blast, the shock wave signature if the projectile is traveling at supersonic speed, and possibly even the characteristic sound of the firearm's mechanical action if the recording is obtained close to the shooting position. To investigate the acoustical phenomena associated with gunshot evidence, a systematic set of rifle shots were made from distances ranging from 10 meters to nearly 800 meters away from the recording microphone. This paper summarizes the primary acoustical evidence derived from these recorded gunshots, and suggests several strengths and weaknesses of gunshot analysis for forensic purposes.
A system was developed able to retrieve specific documents from a document collection. In this system the query is given in text by the user and then transformed into image. Appropriate features were in order to captu...
详细信息
Wireless sensor networks (WSNs) have been of great interest among academia and industry due to their diverse applications in recent years. The main goal of a WSN is data collection. As the amount of the collected data...
详细信息
This paper introduces a new signature coding which is designed based on the well-known Block Truncation Coding (BTC). It comprises of bit-maps of the signature blocks generated by different threshold criteria. Two new...
详细信息
ISBN:
(纸本)9780819473066
This paper introduces a new signature coding which is designed based on the well-known Block Truncation Coding (BTC). It comprises of bit-maps of the signature blocks generated by different threshold criteria. Two new BTC-based algorithms are developed for signature coding, to be called Block Truncation Signature Coding (BTSC) and 2-level BTSC (2BTSC). In order to compare the developed BTC based algorithms with current binary signature coding schemes such as Spectral Program Analysis Manager (SPAM) developed by Mazer et al. and Spectral Feature-based Binary Coding (SFBC) by Qian et al, three different thresholding functions, local block mean, local block gradient, local block correlation are derived to improve the BTSC performance where the combined bit-maps generated by these thresholds can provide better spectral signature characterization. Experimental results reveal that the new BTC-based signature coding performs more effectively in characterizing spectral variations than currently available binary signature coding methods.
Wireless sensor networks (WSNs) have attracted much interest in recent years. The main goal of a WSN is data collection. As the amount of the collected data increases, it would be essential to analyze them. However, r...
详细信息
Wireless sensor networks (WSNs) have attracted much interest in recent years. The main goal of a WSN is data collection. As the amount of the collected data increases, it would be essential to analyze them. However, restricted power supply of small sensors makes it a serious problem to transmit all the data to a fusion center for a centralized analysis. This is why the role of in-network processing seems crucial. In this paper, we propose an in-network optimization algorithm based on Nelder-Mead simplex to incrementally do regression analysis over distributed data. Then, we improve the resulted regressor by the application of boosting concept. Simulation results show that the proposed algorithm increases accuracy and is more efficient in terms of communication compared to its counterparts.
The positive results seen by ES cells precociously expressing the vWF and PECAM genetic markers on the tetraglyme surfaces suggest a directed differentiation of ES cells into endothelial cells. Surfaces with more cros...
详细信息
ISBN:
(纸本)9781615670802
The positive results seen by ES cells precociously expressing the vWF and PECAM genetic markers on the tetraglyme surfaces suggest a directed differentiation of ES cells into endothelial cells. Surfaces with more cross-linking and higher C:O ratios appear to enhance this differentiation pattern.
Document binarization is an active research area for many years. There are many difficulties associated with satisfactory binarization of document images and especially in cases of degraded historical documents. In th...
详细信息
Historically, for purposes of calibration, the Landsat 4 Thematic Mapper (TM) sensor has relied upon prelaunch calibration data, internal calibration data and very limited post-launch data. Landsat 4 additionally has ...
详细信息
ISBN:
(纸本)9781424428083
Historically, for purposes of calibration, the Landsat 4 Thematic Mapper (TM) sensor has relied upon prelaunch calibration data, internal calibration data and very limited post-launch data. Landsat 4 additionally has used Landsat 5 TM cross-calibration data. The availability of one more independent source, commonly called pseudo-invariant sites, can be used to improve the calibration of the sensor. Pseudo-invariant sites are mainly well characterized desert sites exhibiting the properties of high reflectance, high spatial and spectral uniformity, temporal stability and are nearly Lambertian to decrease BRDF and shadow effects. The successful update of the calibration of the Landsat 5 TM sensor using the pseudo-invariant site approach suggested using this approach to calibrate the Landsat 4 TM sensor [1]. The pseudo-invariant sites selected to characterize the Landsat 4 TM sensor are Middle East sites at Path 166, Row 39 and Path 166, Row 40. Previously, calibration attempts have been primarily based only on the internal calibrator (IC) and prelaunch data. The Landsat 5 internal calibrator was found to be unreliable approximately three years after launch. The Landsat 4 IC data also become questionable especially knowing that the internal calibrators for both the instruments were essentially identical. Thus, it is necessary to characterize the system with another independent and reliable source-in this case pseudo-invariant sites. Unfortunately, due to the unavailability of an adequate number of scenes at the site used for Landsat 5, the Middle East site of Path 166, Row 39 and Path 166, Row 40 were selected. A large drawback of these sites is that they are the locations of the First Gulf War in 1991. This negates the assumption that the surface and atmosphere did not change during the study period. To account for the change, radiometric correction was done using the known Landsat 5 gain signature. A correction factor, which is the ratio of the averages of Landsat 5 p
暂无评论