A Modified Speckle Suppression Algorithm for Breast Ultrasound Images Using Directional Filters

Despeckling of ultrasound images using non-linear conductance function

“…The performance of filtering depends on the diffusion coefficient c which means that the diffusion process can be controlled by optimally selecting c. The basic idea is to make it approach towards unity in the interior of a region (to enhance diffusion in the regions whose pixel values have lesser difference) and towardszero at the boundaries (to stop diffusion at regions with greater difference of pixel values between neighbors thereby avoiding blurring) [24]- [25]. Homogenous regions are considered to be devoid of any edges or boundaries, so in these areas, an edge seeking function is selected which is more of a 'filtering' function than an 'edge-preserving' function.…”

Speckle suppression in SAR images employing modified anisotropic diffusion filtering in wavelet domain for environment monitoring

“…The difference between E 2 and E 1 is therefore computed: (16) where: ∆E serves to define the stopping criterion. If ∆E ≥ 0, then the process is terminated else I 2 (x, y) is again replaced by the filtered image I 3 (x, y) obtained at the next filtering iteration; E 2 is stored in E 1 and new error E 2 is generated using I 3 (x, y) and I 2 (x, y) in Eq.…”

Bilateral despeckling filter in homogeneity domain for breast ultrasound images

“…After this, using a suitable threshold, the complete image is classified into two regions-homogeneous and nonhomogeneous. Homogeneous regions contain the inner parts of the region (intra-region information) while the nonhomogeneous regions contain the edges and boundaries (interregion information) [2]. The procedural steps involved are mentioned below:…”

“…The utility of ultrasonography for clinical assessment and diagnosis of breast cancer is well established owing to its noninvasive nature (not requiring any ionizing radiations) and low cost [1]- [2]. Although, continued improvement in ultrasound image quality is equally desirable and becomes a challenge in accurate assessment of these images.…”

“…Filtering approaches in homogeneity domain include homog filters which employ operations performing histogram improvement, in which dominant bright pixels are de-emphasized [24]- [25]. Applicability of directional filters to non-homogeneous regions of medical images yielded convincing results in work of et al [26], [2]. Although, there existed certain issues related to residual speckle content and reduced contrast in despeckled images.…”