Posts by SWe

Dear Sangrae,

thank you for your reply. I'm glad to hear from your positive results.

You're welcome.

Best regards,

Simon

Dear Sangrae_Kim,

this is a real basic approach, maybe it would be usefull to use a

small binomial filter (x-direction) after the box filtering to avoid the sinc-artifacts.

I'm really curious about your results.

Best regards,

Simon

Dear Sangrae_Kim,

I made you a simple prototype design for line lengths of 1024 pixel.

Inside of "FrequencyBoxFilter" you can specify two ranges of frequencies to remove - as you need to remove the positive and the negative frequencies.

I hope this helps a little bit.

Greetings,

Simon

Hello,

when you have the need to transfer small images over the DMA, the transfer bandwidth decreases.

If you would concatenate a series of images for the transfer you would overcome this. The downside would be that you only would get a new image in the host system when a series is complete.

Problem example:

You transfer small ROIs of BLOBs with a speed of several thousands of FPS. You concatenate 10 of these images to increase the DMA efficiency, but you only have 42 BLOBs in your frame. You would get 4 "big" frames over the DMA, then you have to wait (maybe a long time) for the next 8 BLOBs to complete the next 10 frame series .

Solution:

To overcome this limitation, a timeout is introduced. As seen in the screenshot below, a SignalGate is used to cut the data stream if the start of the next frame takes too long. To achieve this functionality, the data stream has to be converted to 1D. The ImageValve in front of the SignalGate ensures that no image data reaches the SignalGate when it's closed.

Limitation:

The input frames should be of constant height.

flush.png

I created a test design for you. Please keep in mind that you can't simulate it. You have to build it and test it with hardware.

HowTo:

1. Change the design to your target platform
2. Build
3. Load the Applet
4. Start acquisition
5. Set signal of "Go" to "High", after a little bit of time, set it to "Low"
6. After some time you should get an additional (incomplete series) frame. Timeout is set to 1,6s on ME5 platforms.

overview.png

Greetings,

Simon

Dear B.Ru,

we are now able to extract the objects themselves. Which now enables the possibility of calculating features on the input data, like the mean gray value and non additive features of the object.

Right now I started implementing the gray value statistics, as they are straightforward to implement.

Best regards

Simon

For other readers: I would like to encourage you to help me with the Questions in #3, having more replies than just one would help me much! Thank you

Hello,

in the past weeks I worked out a list of features in binary, gray and color domain. For clarification: they are not implemented in VA so far (except the features obtained by the BLOB operators)!

I would appreciate further help from experienced people in object classification, there are the following questions:

1. If you could choose 10 of the listed features, which of these would it be?
2. Do I miss important Features?

If you want to help me, please use the attached .pdf, annotate it and post it here or send it via e-mail to me (address on request).

Thanks in advance

Simon

Featurelist_Forum.pdf

Hello,

I am currently working on my master thesis at Silicon Software. I would like to ask you for help!

The thesis is about BLOB based object classification. The main part is to select the most important parameters like shape factors, etc. and implement their calculation in the FPGA via VA.

So I would like to ask you two questions:

1. What do you think are the most important BLOB parameters to use for object classification?
2. Do you have example images, maybe of field applications, for the usage of the parameters mentioned in 1)?.

Thanks in advance

Simon