Part Number: TDA2
Hi,
I am a beginner and I just started to work on TDA2EVM and Adaboost for OD application. I want to train the Adaboost classifier for Pedestrian detection which should execute on TDA2.
From the previous discussions about AdaBoost training on this forum, I got to know that we can use Adaboost object detection training tool (Matlab based github.com/.../acf-jacinto ) to generate .descriptor.
I read the info in the Readme file
## Usage * Open Matlab and navigate to detector folder. * Open acfJacintoExample.m in editor * Make changes for your dataset path, list of videos and annotations files, object type to be trained etc. * Run the file to do train and test.
I want to do the Adaboost training with Caltech dataset for pedestrian detection---- > http://www.vision.caltech.edu/Image_Datasets/CaltechPedestrians/
Has anyone tried training and testing the Adaboost classifier with Caltech dataset?
So In acfJacintoExample.m I should just give the respective paths to .seq and respective annotations .vbb file and then execute the .m?
elseif strcmp(dataName, 'CaltechUsa')
%error('Define the dataset paths here');
exptName='AcfJacintoClatechUsa';
extractType='all';
extractFormat='';
dataDir='E:\HOG_SVM\Training_datasets_SVM\Calltech_for_Adaboost';
vidList={ ...
{'videos/set00/V000.seq', 'videos/set00/V001.seq','videos/set00/V002.seq', 'videos/set00/V003.seq','videos/set00/V004.seq', 'videos/set00/V005.seq'}, ...
{'videos/set01/V000.seq', 'videos/set01/V001.seq','videos/set01/V002.seq', 'videos/set01/V003.seq','videos/set01/V004.seq', 'videos/set01/V005.seq'} ...
};
vbbList={ ...
{'annotations/set00/V000.vbb', 'annotations/set00/V001.vbb', 'annotations/set00/V002.vbb', 'annotations/set00/V003.vbb', 'annotations/set00/V004.vbb', 'annotations/set00/V005.vbb'}, ...
{'annotations/set01/V000.vbb', 'annotations/set01/V001.vbb', 'annotations/set01/V002.vbb', 'annotations/set01/V003.vbb', 'annotations/set01/V004.vbb', 'annotations/set01/V005.vbb'} ...
};
config = [];
-------------------------------------------------------------------------------------------------------------------------------------------------
Do I need to do anything else like setting up any parameters before training and executing acfJacintoExample.m ??
-------------------------------------------------------------------------------------------------------------------------------------------------
I got to know that after execution know that a .descriptor file will be generated which we need to convert into binary format which is TDA OD application compatiable weight file by using AdaboostTableGen.exe. Anything else I should know more like type conversions of some variables or so?





















