Limits...
Microscopic images dataset for automation of RBCs counting.

Abbas S - Data Brief (2015)

Bottom Line: The Dataset consists of Red Blood Corpuscles (RBCs) images and there RBCs segmented images.A detailed description using flow chart is given in order to show how to produce RBCs mask.The RBCs mask was used to count the number of RBCs in the blood smear image.

View Article: PubMed Central - PubMed

Affiliation: Ain Shams University, Cairo, Egypt ; Middle East Technical University, Ankara, Turkey.

ABSTRACT
A method for Red Blood Corpuscles (RBCs) counting has been developed using RBCs light microscopic images and Matlab algorithm. The Dataset consists of Red Blood Corpuscles (RBCs) images and there RBCs segmented images. A detailed description using flow chart is given in order to show how to produce RBCs mask. The RBCs mask was used to count the number of RBCs in the blood smear image.

No MeSH data available.


Initial segmentation.
© Copyright Policy - CC BY
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4556816&req=5

f0015: Initial segmentation.

Mentions: The acquired images subjected to segmentation process as a part of the date reduction stage involves the partitioning of the image plane into meaningful parts [1,2]. In this work a system scheme (Diagram 1) is proposed to segment both the RBCs form the background and the counting area which indicated by the triple line square on the glass slide [3]. The full description of the segmentation scheme is given as follows: At the first step, the captured smear image split into its three color component bands (red, green and blue) as shown in Fig. 1. The three grayscale components are examined and we note that the green components show a clear image of RBCs [4]. At the second step, the histogram green component has been performed and the initial threshold value taken as gray level corresponding to the peak of the histogram as shown in Fig. 2. At the third step, the areas of the objects on the threshold image has been determined which shows that the most common areas are less than 10 pixels corresponding to the area of the RBCs. The objects that have areas larger than this value (10) are due to inaccurate threshold as shown in Fig. 3. Based on this, if there is any object which has an area greater than 10 the feedback increases the value of threshold by one until no longer object has an area greater than 10 as shown in Fig. 4. At the fourth step, extracting the counting area by calibrating the cropped image to be the same area of the triple line square at that system setup as shown in Fig. 5. Finally, the number of objects in the extracted area determined which is corresponding to the number of RBCs in that area. A complete set of images, their RBCs counts and segmentations are included as supplementary material to this article.


Microscopic images dataset for automation of RBCs counting.

Abbas S - Data Brief (2015)

Initial segmentation.
© Copyright Policy - CC BY
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4556816&req=5

f0015: Initial segmentation.
Mentions: The acquired images subjected to segmentation process as a part of the date reduction stage involves the partitioning of the image plane into meaningful parts [1,2]. In this work a system scheme (Diagram 1) is proposed to segment both the RBCs form the background and the counting area which indicated by the triple line square on the glass slide [3]. The full description of the segmentation scheme is given as follows: At the first step, the captured smear image split into its three color component bands (red, green and blue) as shown in Fig. 1. The three grayscale components are examined and we note that the green components show a clear image of RBCs [4]. At the second step, the histogram green component has been performed and the initial threshold value taken as gray level corresponding to the peak of the histogram as shown in Fig. 2. At the third step, the areas of the objects on the threshold image has been determined which shows that the most common areas are less than 10 pixels corresponding to the area of the RBCs. The objects that have areas larger than this value (10) are due to inaccurate threshold as shown in Fig. 3. Based on this, if there is any object which has an area greater than 10 the feedback increases the value of threshold by one until no longer object has an area greater than 10 as shown in Fig. 4. At the fourth step, extracting the counting area by calibrating the cropped image to be the same area of the triple line square at that system setup as shown in Fig. 5. Finally, the number of objects in the extracted area determined which is corresponding to the number of RBCs in that area. A complete set of images, their RBCs counts and segmentations are included as supplementary material to this article.

Bottom Line: The Dataset consists of Red Blood Corpuscles (RBCs) images and there RBCs segmented images.A detailed description using flow chart is given in order to show how to produce RBCs mask.The RBCs mask was used to count the number of RBCs in the blood smear image.

View Article: PubMed Central - PubMed

Affiliation: Ain Shams University, Cairo, Egypt ; Middle East Technical University, Ankara, Turkey.

ABSTRACT
A method for Red Blood Corpuscles (RBCs) counting has been developed using RBCs light microscopic images and Matlab algorithm. The Dataset consists of Red Blood Corpuscles (RBCs) images and there RBCs segmented images. A detailed description using flow chart is given in order to show how to produce RBCs mask. The RBCs mask was used to count the number of RBCs in the blood smear image.

No MeSH data available.