Electronic medical imaging technologies are growing rapidly and simplifying diagnosis in medical ... more Electronic medical imaging technologies are growing rapidly and simplifying diagnosis in medical area. The proper use of this technology requires a better understanding, interpretation, and development of new, efficient algorithms. Processing and recognition techniques of patterns related to these medical devices are becoming more important. Among these techniques artificial neural network structures are very promising in the diagnosis decision support mechanisms. In this paper, it is aimed to present the performance of statistical neural network structures on classifying cardiac problems which are obtained from SPECT (Single Photon Emission Computed Tomography) images. Principal component analysis has been used to overcome excessive dimensionality of data. After classification we used Receiver Operation Characteristics (ROC) analysis to evaluate system performance. Results show that proper neural network based statistical pattern recognition models will play a fundamental role in medical signal processing and image analysis.
... Oguz Altun, Songul Albayrak Yildiz Teknik Universitesi Bilgisayar Muihendisligi B6lumui {oguz... more ... Oguz Altun, Songul Albayrak Yildiz Teknik Universitesi Bilgisayar Muihendisligi B6lumui {oguz, songul}@ce.yildiz.edu.tr Ozetce ... Kalin kenar ,ikarimi yontemi el ,ekillerinin dahaince kisimlarinin (parmaklar gibi) sonuca etkisini artirir. ...
Fingerspelling is used in sign language to spell out names of people and places for which there i... more Fingerspelling is used in sign language to spell out names of people and places for which there is no sign or for which the sign is not known. In this work we describe a method for increasing the effect of fingers in Fingerspelling hand shapes. Hand shape objects are obtained by extraction of representative frames, color segmentation in YCrCb space and angle of least inertia based fast alignment [1]. Thick edges of the hand shape objects are extracted with a distance to edge based method. Finally a calculation that penalizes similarity for not-corresponding pixels is employed to correlation based template matching. The experimental Turkish fingerspelling recognition system recognizes all 29 letters of the Turkish alphabet. The train video database is created by three signers, and has a set of 290 videos. The test video database is created by four signers, and has a set of 203 videos. Our methods achieve a success rate of 99%.
Fingerspelling is used in sign language to spell out names of people and places for which there i... more Fingerspelling is used in sign language to spell out names of people and places for which there is no sign or for which the sign is not known. In this work we describe a Turkish fingerspelling recognition system that recognizes all 29 letters of the Turkish alphabet. A single representative frame is extracted from the sign video, since that frame is enough for recognition purposes of the letters mentioned. Processing a single frame, instead of the whole video, increases speed considerably. The skin regions in the representative frame are extracted by color segmentation in YCrCb space before clearing noise regions by morphological opening. A novel fast alignment method that uses the angle of orientation between the axis of least inertia and y axis is applied to hand regions. This method compensates small orientation differences but increases big ones. This is desirable when differentiating the fingerspelling signs, some of which are close in shape but different in orientation. Also the use of minimum bounding square is advised, which helps in resizing without breaking the alignment. Binary values of this minimum bounding square are directly used as feature values, and that allowed experimenting with different classification schemes. Features like mean radial distance and circularity are also used for increasing success rate. Classifiers like kNN, SVM, Naïve Bayes, and RBF Network are experimented with, and 1NN and SVM are found to be the best two of them. The video database was created by 3 different signers, a set of 290 training videos, and a separate set of 174 testing videos are used in experiments. The best classifiers 1NN and SVM achieved a success rate of 99.43% and 98.83% respectively.
Electronic medical imaging technologies are growing rapidly and simplifying diagnosis in medical ... more Electronic medical imaging technologies are growing rapidly and simplifying diagnosis in medical area. The proper use of this technology requires a better understanding, interpretation, and development of new, efficient algorithms. Processing and recognition techniques of patterns related to these medical devices are becoming more important. Among these techniques artificial neural network structures are very promising in the diagnosis decision support mechanisms. In this paper, it is aimed to present the performance of statistical neural network structures on classifying cardiac problems which are obtained from SPECT (Single Photon Emission Computed Tomography) images. Principal component analysis has been used to overcome excessive dimensionality of data. After classification we used Receiver Operation Characteristics (ROC) analysis to evaluate system performance. Results show that proper neural network based statistical pattern recognition models will play a fundamental role in medical signal processing and image analysis.
... Oguz Altun, Songul Albayrak Yildiz Teknik Universitesi Bilgisayar Muihendisligi B6lumui {oguz... more ... Oguz Altun, Songul Albayrak Yildiz Teknik Universitesi Bilgisayar Muihendisligi B6lumui {oguz, songul}@ce.yildiz.edu.tr Ozetce ... Kalin kenar ,ikarimi yontemi el ,ekillerinin dahaince kisimlarinin (parmaklar gibi) sonuca etkisini artirir. ...
Fingerspelling is used in sign language to spell out names of people and places for which there i... more Fingerspelling is used in sign language to spell out names of people and places for which there is no sign or for which the sign is not known. In this work we describe a method for increasing the effect of fingers in Fingerspelling hand shapes. Hand shape objects are obtained by extraction of representative frames, color segmentation in YCrCb space and angle of least inertia based fast alignment [1]. Thick edges of the hand shape objects are extracted with a distance to edge based method. Finally a calculation that penalizes similarity for not-corresponding pixels is employed to correlation based template matching. The experimental Turkish fingerspelling recognition system recognizes all 29 letters of the Turkish alphabet. The train video database is created by three signers, and has a set of 290 videos. The test video database is created by four signers, and has a set of 203 videos. Our methods achieve a success rate of 99%.
Fingerspelling is used in sign language to spell out names of people and places for which there i... more Fingerspelling is used in sign language to spell out names of people and places for which there is no sign or for which the sign is not known. In this work we describe a Turkish fingerspelling recognition system that recognizes all 29 letters of the Turkish alphabet. A single representative frame is extracted from the sign video, since that frame is enough for recognition purposes of the letters mentioned. Processing a single frame, instead of the whole video, increases speed considerably. The skin regions in the representative frame are extracted by color segmentation in YCrCb space before clearing noise regions by morphological opening. A novel fast alignment method that uses the angle of orientation between the axis of least inertia and y axis is applied to hand regions. This method compensates small orientation differences but increases big ones. This is desirable when differentiating the fingerspelling signs, some of which are close in shape but different in orientation. Also the use of minimum bounding square is advised, which helps in resizing without breaking the alignment. Binary values of this minimum bounding square are directly used as feature values, and that allowed experimenting with different classification schemes. Features like mean radial distance and circularity are also used for increasing success rate. Classifiers like kNN, SVM, Naïve Bayes, and RBF Network are experimented with, and 1NN and SVM are found to be the best two of them. The video database was created by 3 different signers, a set of 290 training videos, and a separate set of 174 testing videos are used in experiments. The best classifiers 1NN and SVM achieved a success rate of 99.43% and 98.83% respectively.
Uploads