Sa sve većom primenom digitalne slike u raznim oblastima nauke javljaju se i novi izazovi za obradu. Nove osobine slike potrebno je valjano obraditi i opisati posmatrajući ih iz više uglova bazirajući se na teorijska razmatranja. Ovaj rad definiše karakteristike Snow & Rain šuma digitalne slike metodom parcijalnog filtriranja i pruža jedan od modela za detektovanje vrste i koncentracije šuma u digitalnoj slici. Rezultati su prikazani grafički i numerički nakon obrade više od 50 digitalnih slika, a tretirani su adekvatnim parametrima za ocenu kvaliteta.

This paper provides edge detection analysis on images, which consist of different numbers of details (small, medium and high number of details) and which are compressed by different compression algorithms -JPEG, JPEG2000 and SPIHT. Images from the BSD (Berkeley Segmentation Database) database were used and compressed with different number of bits per pixel. The analysis was performed for five edge detectors: Canny, LoG, Sobel, Prewitt, and Roberts. The fidelity of the detected edges was determined using the objective measures Figure of Merit (FOM), F measure and Performance Ratio (PR), where the reference value was taken from the GroundTruth image. Based on the results presented in the tables, it can be concluded that edge detection behaves differently depending on the number of bits per pixel and applied compression algorithm, as well as, the number of details in the image. Roberts operator has been proven to be the best solution, when it is necessary to perform better edge detection over compressed images with small a number of details, but Canny shows better results for images with a high number of details.

Image segmentation is the process of dividing a digital image into image segments so that individual regions of interest can be analyzed and processed instead of the entire image. Image segmentation has a significant role in detecting regions of interest and extracting attributes and regions from those images. In this paper, five original grayscale abnormal MRI brain images have been processed by using image segmentation techniques for detecting and extracting regions of interest, in this case, tumors. This research described three methods of detection and extraction of tumors from abnormal MRI brain images in MATLAB: a method based on combined local threshold segmentation techniques with morphological operations for tumor detection; a method based on region splitting and merging segmentation techniques; and a method based on combined thresholding, Meyer's flooding watershed algorithm, as an image segmentation technique with morphological operations for tumor detection. Abnormal MRI brain images were preprocessed in order to obtain suitable results. Image data used in this research were obtained from Radiopedia, an educational radiology resource. The best method for detecting and extracting tumors has been determined by comparing the results of accuracy, sensitivity, F-measure, precision, MCC, dice, jaccard, and specificity. Based on the results of these measurements, it has been concluded and confirmed that the first and third methods are both equally good for detecting and extracting tumors.