/Tensorflow_API_object_detection

Primary LanguageJupyter Notebook

Rock-Paper-Scissor object_detection using tensorflow API

Overview

This project is an object detection and localization using Object Detection API with TensorFlow 2 where we used a dataset of three classes: Rock, Paper, and Scissor. We annotated and labeled the images to prepare for training, then used two different pre-trained models to compare performances. Models used are: SSD MobileNet_v2_320x320 and efficientdet_d0_coco17_tpu_32, both models were trained for 2000 epochs.

Code:

Notebooks are divided into two sections:

  1. Data preprocessing & training

  2. Evaluation