1 / 17

Is olating Objects From Image Stack

Is olating Objects From Image Stack. Presented By: Md. Amjad Hossain and Raja Naresh. Objectives. Constracting objects from the image sequence constructed by slicing the objects. Objectives. Constructing objects from the image sequence obtained by slicing the objects. A Problem Instance.

ashtyn
Télécharger la présentation

Is olating Objects From Image Stack

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Isolating Objects From Image Stack Presented By: Md. AmjadHossain and Raja Naresh

  2. Objectives • Constracting objects from the image sequence constructed by slicing the objects.

  3. Objectives • Constructing objects from the image sequence obtained by slicing the objects.

  4. A Problem Instance Isolating a single Axon from the stack of images. • Identify each part of same axon in each image of the stack. • Merge all parts to get the 3D form of the axon.

  5. A Problem Instance Isolating a single Axon from the stack of images. • Identify if the axon is splitting from one image to another image.

  6. A Problem Instance Isolating a single Axon from the stack of images. • Identify if two or more axon’s parts are getting merged from one image to another image.

  7. Region Growing algorithm - Used to identify a single part of axon within an image. Region Growing (SeedPoint, Threshold). 1. insert SeedPoint into queue 2. Calculate low = SeedPoint<Threshold? 0 : Threshold High = SeedPoint<Threshold? Threshold : 255 3. while(queue not empty) if (neighbour >= low && neighbour <= high) insert neighbour into queue. include neighbour in the axon set.

  8. Algorithm • Identifying seed point and region growing for the consecutive image in the stack. Scan through the mirroring coordinates in the second image of the stack and identify a pixel between low and high calculated from threshold. image 1 Pixel is not related to initial seed point image 2 image 1 image 2

  9. Algorithm • Identifying seed point and region growing for the consecutive image in the stack. Splitting case. Repeat the seed identification process if a significant number of pixels are still left. The above process will allow us to detect any splits occurring in the consequent image slices in the stack.

  10. Algorithm • Identifying seed point and region growing for the consecutive image in the stack. Merging case. • After getting all parts of the axon in an image check whether those parts are branching in the previous images. - Spotting split in reverse stack is basically merge.

  11. Full object detection algorithm

  12. Full object detection algorithm

  13. Result-1 ( Isolating a single axon) To isolate object, Load image sequence Apply filtering (Gaussian Blur) Select initial seed point 5. Apply thresholding to separate object pixels from other pixels. 6. Use 3D viewer to see the object 4. Apply the object detection algorithm

  14. Result-2 ( Isolating multiple axons) The steps are same as isolating single axon. Select multiple seed points (one on each axon) on the first image of the image stack. Initial Image Stack 3D view Isolated axons

  15. Result-3 Isolating all connected neurons from confocal image. Select one or multiple seed points In this case, one seed point has been selected from 13th image of the original image stack

  16. Comparison with existing fast marching • Fast marching in Fiji always convert original image to low quality image by applying a fixed filter. But the new algorithm can be applied to images of any quality. • Although the Fast marching technique uses lower quality image, the process is extremely slow while our algorithm is very faster.

  17. Conclusion and Future work - We implemented very basic but crucial part for isolating objects from image stack very quickly. - We have figure out algorithm for calculating surface area of the object detected but didn’t implemented as it is very trivial algorithm. - As we didn’t know the actual voxel height so we didn’t go through implementation of calculating volume of the object. The work can be extended to • Calculating threshold value automatically. • Calculate approximate surface area, volume of the objects detected

More Related