No description available

A pure Tensorflow+Keras TPU trainable implementation of SSD (Single Shot MultiBox Detector) using different backbones of EfficientNet which can be replaced with any ImageNet backbone. Tested on the PASCAL_VOC dataset. Implemented several low level data preprocessing and augmentations in pure tf functions for faster computation.

This is the object detection model trained on Pascal VOC 2012 dataset using the Tensorflow's object detection API. A new model is trained for 20 classes using the ssd_mobilenet_v1

🏠 End-to-end TensorFlow 1.x pipeline for training and validating SSD MobileNet V2 object detection on custom PASCAL-VOC data, with Colab notebooks for dataset prep, training, and inference.

Object detection practice project using TensorFlow and SSD MobileNet V2 on the pascal VOC 2007 dataset. This project explores real_time object detection, model evaluation, and performance analysis using metrics like IOU,percision, and recall.