Android Tensorflow lite custom object detection

preview_player
Показать описание
This is a custom object detector trained with Tensorflow lite for android...It has only 6 objects as the cards nine,ten,jack,queen,king,ace...I trained this model using ssd_mobilenet_v2_quantized_300x300_coco on tensorflow 1.15 and bazel 0.26.1 to convert it to tflite on linux as i faced issued building bazel on windows....Let me know if anyone wants a tutorial video...
App building link
Detect and labelimages link
App link
Рекомендации по теме
Комментарии
Автор

Working perfectly... Kudos for your work 💕❤️

roshanshaji
Автор

Hey great work.I am trying to achieve this since months but not able to achieve I make tflite models successful but nothing is captured on android, no bounding boxes are visible . Can you please guide me .It would be really kind of you.

karuneshpalekar
Автор

Please do let me know if anyone is having issues with the app

doomsdayrobotics
Автор

I want model which detect all 13 cards, not only 6. Can you help me

muhammadahmadraza
Автор

Please make a tutorial video or a article of how you made this possible.I am really on the stage of giving up.I have wasted like 2 months on making it work, but Ihave failed multiple times

karuneshpalekar
Автор

I have my own data(images) but i dont know how to create tflite model. Can you help me please?

hamiduzzamannoyon
join shbcf.ru