Fusing the line intensity profile and support vector machine for removing reflections in frontal RGB color eye images

A low lighting or contrast ratio visible iris recognition using iso-contrast limited adaptive histogram equalization

“…Tables 5-9 present the results of the contrast enhancement analyses for different categories of eye images, methods, contrast levels and distances. As seen in the results, the intensity of the eye images from the Raffei et al [21] method was increased, as well as the RMS values, following the implementation of the contrast enhancement methods. For example, the RMS values for the eye images obtained from the Raffei et al [21] method at a distance of 4 m ( Table 5) increased from the low level of contrast to …”

“…The reflection removal and proposed contrast enhancement methods are implemented in the preprocessing stage. Description for each of the stages can be found in Raffei et al [21] except for the contrast enhancement. As discussed in the introduction, the three issues that reduce iris recognition accuracy are the eye image with a low lighting, the eye image with a low contrast ratio between the iris and pupil, and the eye image with both low lighting and low contrast ratio between the iris and pupil.…”

“…Several evaluations and analyses on the 1000 eye images were completed in order to compare the performance of the proposed iris recognition system and the existing methods proposed by Raffei et al [21], Santos and Hoyle [25], Shin et al [26], and Shukri et al [27]. It should be noted that for satisfactory comparisons, the outputs of the reflection removal eye images from the method developed by Raffei et al [21] were used as the input images.…”

“…It should be noted that for satisfactory comparisons, the outputs of the reflection removal eye images from the method developed by Raffei et al [21] were used as the input images. The most common measures of image quality in image processing: intensity, namely peak signal-to noise ratio (PSNR), mean square error (MSE) and absolute mean brightness error (AMBE) were used to compare the performance of the contrast enhancement methods [8].…”

“…Reflections can also be observed in the eye image with low lighting after the process of contrast enhancement. Therefore, a combination of the line intensity profile and support vector machine (LIPSVM) proposed by Raffei et al [21] is applied to the eye image to remove the reflections and obtain the correct iris boundaries. A summary of the proposed contrast enhancement process is illustrated in Fig.…”