CCS/TDA2EVM5777: Please explain the source code for the semantic segmentation usecase

user6187828

Part Number: TDA2EVM5777

Tool/software: Code Composer Studio

I am refering the tidl usecase so for that i wnt to create the tidl semantic segmentation usecase for that i need some guideline

over 4 years ago

0 Gaviraju Doddabettahalli Bettegowda over 4 years ago

TI__Genius 14305 points

Hi,

Please refer chapter 4 in this document, vision_sdk/docs/FeatureSpecificUserGuides/VisionSDK_UserGuide_TI_DeepLearning.pdf

segmentation usecase is available in the below folder.

vision_sdk\apps\src\rtos\usecases\semSeg

Please refer chains_semSeg.txt file for all the links used in this usecase.

Can you be a bit more specific about information you want?

Regards

Gaviraju

0 user6188127 over 4 years ago in reply to Gaviraju Doddabettahalli Bettegowda

Prodigy 210 points

Sir I am not understanding the source code for the semantic segmentation usecase

0 Gaviraju Doddabettahalli Bettegowda over 4 years ago in reply to user6188127

TI__Genius 14305 points

Do you understand Vision sdk links and chain architecture?

Are you facing issue in understanding the usecase source code file available in below folder

vision_sdk\apps\src\rtos\usecases\semSeg

Or you want to know how tidl algos are working and how road sign or vehicles are detected?

Thanks

Gaviraju

0 user6188127 over 4 years ago in reply to Gaviraju Doddabettahalli Bettegowda

Prodigy 210 points

No sir I am not able to understanding the links and chains architecture and when we creating the usecase how they are using the different links and so that's why I am getting

0 Gaviraju Doddabettahalli Bettegowda over 4 years ago in reply to user6188127

TI__Genius 14305 points

Hi,

Please refer below document for understanding the link & chain architecture of VSDK.

vision_sdk/docs/Architecture/VisionSDK_SW_Architecture_Overview.pdf

vision_sdk/docs/Architecture/VisionSDK_SW_Architecture_Details.pdf

Auto Usecase Generation tool - Usage, tool enhancements to support new links refer to this: vision_sdk/docs/FeatureSpecificUserGuides/VisionSDK_UserGuide_UsecaseGen.pdf

vision_sdk/docs/FeatureSpecificUserGuides/VisionSDK_Overview_UsecaseGen.pdf

Build system – details of usage, fast builds, custom build configuration

vision_sdk/docs/FeatureSpecificUserGuides/VisionSDK_UserGuide_BuildSystem.pdf

VSDK development guide: vision_sdk/docs/FeatureSpecificUserGuides/VisionSDK_DevelopmentGuide.pdf

Thanks

Gaviraju

0 user6188127 over 4 years ago in reply to Gaviraju Doddabettahalli Bettegowda

Prodigy 210 points

For understanding the tidlpreproc and tidlpostproc algorithm please explain the this

0 Gaviraju Doddabettahalli Bettegowda over 4 years ago in reply to user6188127

TI__Genius 14305 points

In tidlpreproc, we are converting a video frame from YUV to BGR data format.

In tidlpostproc, we are creating overly of different objects

Thanks

Gaviraju

0 user6187828 over 4 years ago in reply to Gaviraju Doddabettahalli Bettegowda

Intellectual 600 points

vision_sdk\apps\src\rtos\usecases\semSeg

Or you want to know how tidl algos are working and how road sign or vehicles are detected?

0 Gaviraju Doddabettahalli Bettegowda over 4 years ago in reply to user6187828

TI__Genius 14305 points

Hi

For understanding "how road sign or vehicles are detected" using TIDL Algorithm.

First, you have to understand TI deep learning(TIDL) Algorithm & TIDL comes as a package in Vision SDK.

So I'm requesting to you create new query thread with component name as "TIDL", then our TIDL expert will help you

Thanks

Gaviraju

Processors

Processors forum

CCS/TDA2EVM5777: Please explain the source code for the semantic segmentation usecase