Object Detection Using TensorFlow API

Introduction

This projects aims at object detection with a very high accuracy using pre-trained light weight SSD-MobileNet V3 architecture. My project describes how object detection models can be trined very easily and accurately using the TEnsorFlow API. Here we have also used OpenCV in order to draw the rectangles around the objects and also to load the pre-trained frozen TensorFlow models.

Challenges :

The main challenge faced by me was getting the config file which I finally found at https://github.com/opencv/opencv/wiki/TensorFlow-Object-Detection-API which is the wiki page of TensorFlow API .

Technologies Used :

OpenCV

Installations

Use the package manager pip to install opencv and matplotlib .

pip install opencv-python

pip install matplotlib

BrainStorming

Image Classification/Recognition

It is the process of classifying the images to the categories they belong to based on their salient features.

Deep learning Algorithms for Image Classification: -AlexNet
-GoogleNet
-MobileNet
-VGGNet

Object Detection

Object Detection specifies the location of multiple objects in a image

-Classification
-Localization

Famous Algorithms

-SSD-MobileNetv2,SSD-MobileNetv3
-YOLO v1,v2

Famous Dataset

- COCO-> 80 classes

Run network in OpenCV

OpenCV needs an extra configuration file to import object detection models from TensorFlow. It's based on a text version of the same serialized graph in protocol buffers format (protobuf).

Use existing config file for your model: MobileNet-SSD v3 (Version:2020_01_14)

Config

Contributions:

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change. Please make sure to update tests as appropriate.