ProCodeCG – BDV – Technical Series #8 – Machine Learning Object Detection with YOLO – 8 April 2019 – Day #3 #4

Standard

We finished the #3 and #4 session today 😉

IMG_9638

We explored both darknet and darkflow object detection today. This is the example of darkflow object video with camera

IMG_9639

IMG_9640

IMG_9645

We tried very hard to make either darknet or darkflow works on each computer

IMG_9642

8dd323d5-ee70-42bf-91ff-669e91c62a41

c1e8557c-5869-4373-bd2c-39cbde1ce80d 11e81d99-eca1-44fb-b0d7-171a160d3dd0

88878aba-cca8-4eae-b3b6-9ffc6c32e5fc d3140504-7599-4b47-bb78-ab7b4936b517

54e1d0c8-b31f-4ed1-a20d-613295cac4ae 9daf7f58-557e-441b-99c4-32d985f25037

aff7019a-dce7-4f82-857f-8468ebe6df70 d3489f72-41f1-4307-85b7-62fa99c71ef9

422d8ccc-fead-43f2-973e-cb38e871a140 512d76ee-74c4-499d-a823-9ce887a7865f

b3a190c0-930a-40af-b3b5-6bb5ab53b7a7 dd691013-1794-41f9-87a1-67e354278f72

59ed6d5c-33f6-4f0e-bd53-8fc8f93dd65e

Some of the results today

 

Screen Shot 2019-04-15 at 11.23.14 Screen Shot 2019-04-15 at 11.23.09

The project today is to create our own dataset of 3 kind of instant noodle: Indomie, Supermie and MiSedap. We had to download the picture, put box around it and label it, make the system learn and then see if it can classify or detect the object

IMG_9644

Screen Shot 2019-04-15 at 14.54.57

At the end of the class we were still figuring out about how to get the new weights to train our dataset. The process took lots of time, we had to continue the rest at home 🙂

 

*Photos and videos are also available here
Most of the photos are courtesy of Telkom Bandung Digital Valley

Advertisements

ProCodeCG – BDV – Technical Series #8 – Machine Learning Object Detection with YOLO – 8 April 2019 – Day #2

Standard

Day #2 of deep learning with YOLO 😉

 

photo6149951813193935070

photo6150062494501152960

Today we will test object detection with darknet and darkflow configuration

photo6150062494501152961

It of course took a lot of effort and resources (in terms of time and computational resources :D)

photo6150062494501152962

Every participants are free to decide darkflow or darknet. But both implementations are expected.

photo6150062494501152963

photo6150062494501152964photo6150034358170396903

photo6150062494501152965

photo6150062494501152966

photo6150034358170396905

photo6150062494501152967

photo6150062494501152970

photo6149951813193935074

photo6149951813193935075

photo6150062494501152972

photo6149951813193935078

IMG_9223

IMG_9224

IMG_9225

Some of the participants use Google Colab so they can make use of the GPU 😉

IMG_9228

IMG_9229

An example of the object detection we learned today 🙂

Screen Shot 2019-04-09 at 15.50.39//embedr.flickr.com/assets/client-code.js

predictions//embedr.flickr.com/assets/client-code.js
 

*Photos and videos are also available here
Most of the photos are courtesy of Telkom Bandung Digital Valley

ProCodeCG – BDV – Technical Series #8 – Machine Learning Object Detection with YOLO – 1 April 2019 – Day #1

Standard

Our training about Deep Learning and YOLO for object detection, Day #1 😉

IMG_8664 IMG_8665

Like usual, the first day of the workshop is about basic theories and then installing saga 🙂

photo6127299760537840124

Our target for today is installing Python, virtualenv, tensorflow, requests, numpy, cython, opencv, and nmv

photo6129897330988656723

but since there are lots of operating systems and configurations, we could go as far as installing virtualenv and tensorflow and for some: hello world in tensorflow

photo6129897330988656722

The complexity of the tasks is shown here 😀 *look at the faces LOL

photo6127299760537840129

photo6129897330988656720 photo6129897330988656717

photo6129897330988656716 photo6129897330988656714

photo6129897330988656713 photo6129897330988656712

photo6129897330988656711 photo6129427852408498384

photo6127299760537840122

Next week we will learn about object detection theories and continue the installations 😉

*Photos and videos are also available here. Most of the photos are courtesy of Telkom Bandung Digital Valley

ProCodeCG Kids Regular Coding Class – 30 Mar 2019

Standard

Coding crowd this Saturday 😉

IMG_8533

IMG_8555 IMG_8556

IMG_8557

Hafiz is working on Plant vs Zombie Game. It is quite difficult and takes more time to develop than the previous ones

IMG_8509

Preparing the sprites

IMG_8505

the assets

IMG_8506

each sprite is carefully designed 🙂

IMG_8528 IMG_8534

the background

IMG_8535

IMG_8554

Testing plant placing

IMG_8560

The objects are now placed neatly in the room 😉

IMG_8561

More objects added 🙂 Nice! Next is to add scoring and spawning the zombie randomly 🙂

IMG_8587

Sam’s work is getting harder and harder to do since the target is to release this app to PlayStore

IMG_8507

We need to think how to manage the screens and the size of the app

IMG_8520

We are discussing the screen flow and the best UI/UX design especially to show the user the progress of memorizing the Juz’Amma

IMG_8523

we just realized that Sam’s app has too many screens and that makes the app size grows. The more the screens, the harder it is to manage

IMG_8524

Capt. Faris had his own app developed with Android Studio, and he shared his architecture to us

IMG_8529

it is time to throw in database 😉 and the trick is to create limited screen, and reload the view from the database

IMG_8530

so Sam has to learn about accessing TinyDB 😀

IMG_8545

It leaves him like this LOL

IMG_8583

Fadhlan, stop eating the candy and start to work! 😀

IMG_8508

We have to work on this: testing the right and left motor plus the controller, and make it navigates

IMG_8519

Working hardddd there you go 😀

IMG_8536

IMG_8537

Capt. Awa helped us testing the robot *and ran away when the robot was activated because it refuse to avoid obstacles LOL but at least we know that button, LED, buzzer, wheels, are working and we just need to check the ultrasonic before adding compass and bluetooth module 😀

IMG_8569

Capt. Faris and his research 😀 He was given a challenge to learn how to prevent shoplifting with camera object detection

IMG_8510 IMG_8525 IMG_8546

Capt. Awa, brought new toys!

IMG_8511

The new stuffs have the same size as a coin!

IMG_8512

wowww, modules for TinyDuino

IMG_8513

IMG_8514

IMG_8515

But we still couldn’t get these little mates to work :(( We suspected the USB-UART driver or probably the port

IMG_8526

So Capt. Awa explored another thing: basic electronics simulation

IMG_8542 IMG_8550

IMG_8553 IMG_8558

and then helping us with the robot

IMG_8570

the robot was going in circles so Capt. Awa helped us switching the polarity 🙂

IMG_8572

Testing is done for today (too many things being hit by the robot LOL)

IMG_8574

Badi’s website is ready to be uploaded to public 🙂 Today Badi will buy domain and hosting and upload it

IMG_8521

We’re searching for the best option, and after we got one, Pavilion19 will have a website 🙂

IMG_8540

Aidia is so excited about her Minecraft Project today 🙂

IMG_8517

She brought lots of books for references 🙂

IMG_8518

IMG_8538

IMG_8559

Aidia does both design and mechanical structures in Minecraft

IMG_8564

IMG_8566

Look at this: she created a locker system below a beautiful building she designed 😉 How nice!!!

IMG_8584

Inside the locker corridor 🙂 Next week we’re going to add button to get a specific locker sent to a specific person 😉

IMG_8585

Al is going to make another animation, yaaay 🙂

IMG_8522 IMG_8527

He’s getting better 🙂

IMG_8541

Al finds references on the internet to get more ideas

IMG_8551

And he created animation for educational purposes like this one 🙂

IMG_8552

IMG_8589

We have a newcomer: welcome Andra 🙂

IMG_8543

Like the others, Andra had to go through the intro to programming

IMG_8549 IMG_8563

One challenge down

IMG_8565

Now another one, and Andra can do it very quickly with lots of excitement 🙂

IMG_8567

IMG_8580

Yogi came with the same problem from the last class: displaying a score that follows the object 🙂

IMG_8547

This one worked for Game Maker Studio 2 but not with version 1 so we tried so many ways and still couldn’t figure it out. Don’t worry Yogi we’ll work it out

IMG_6283

Syadad is now ready for a code combat 😉

IMG_8562

Syadad is very excited with this challenge, especially because now he has a friend: Andra 🙂

IMG_8568

Next week Syadad will be ready for creating games yaay!!!

IMG_8581

Until then! 😉

*Photos and videos are available here

ProCodeCG – DevCember 2018 – 22 Dec 2018

Standard

Today we’re invited here to speak about Machine Learning and IoT 🙂

ProCodeCG Kids wait for their turn 🙂

IMG_1712

We gave a demo about Object Detection with Tensorflow to count objects 🙂

Screen Shot 2019-01-19 at 19.58.22

and example of counting objects from a few days before

Screen Shot 2019-01-19 at 19.58.30

and then we took the picture of the audience

IMG_1714

and count the people and it worked 😉

Screen Shot 2019-01-19 at 19.58.33

Us after the presentation 🙂 Thanks for joining the short class 🙂

IMG_1716

Now it’s ProCodeCG Kids turn 🙂

Fadhlan with his brother: Farshad 🙂

IMG_1720 IMG_1721

IMG_1723

Preparing the slides while babysitting 😀

IMG_1726

IMG_1738

IMG_1739

Capt. Awa preparing the presentation

IMG_1722

IMG_1729

This cat feeder will be activated via wifi 🙂

IMG_1724

Nares preparing the slides 🙂

IMG_1725

IMG_1727

The whole family is there to support the kids 😉

IMG_1728

Now the presentation 😉 Capt. Awa

IMG_1740 IMG_1742

IMG_1743

The audience 😉

IMG_1744

IMG_1752

Fadhlan

IMG_1745 IMG_1747

Nares

IMG_1748 IMG_1750

The audience answered our questions correctly got a gift from Traveloka 🙂

IMG_1753

IMG_1754

Closing speech 🙂

IMG_1757

IMG_1758

So proud to be able to be speakers on this awesome event 🙂

IMG_1759

*Photos and videos are also available here

ProCodeCG – BDV – Machine Learning with Python Workshop – Session #3 & #4 – 15 Aug 2018

Standard

Today we’re ready to do some real object detection with TensorFlow 😉

It took us hours and great effort to do installations and debugging etc but we finally made it 😉

IMG_2600 IMG_2601

IMG_2603 IMG_2604

IMG_2605 IMG_2606

IMG_2607

We learned how to do object detection using our own data, using a pre-trained model of course 😉

This is the results 😉

IMG_2609

 

object_detection_workshop1 object_detection_workshop2

object_detection_workshop3 object_detection_workshop4

object_detection_workshop5 object_detection_workshop6

object_detection_workshop7

The system can detect and count the objects 🙂

Bildschirmfoto 2018-08-15 um 7.06.30 PM

Surprised that there are so much staying with us until the end of the class. Thanks for joining the class 🙂

IMG_2634 1