shashankhalo7 / Flipkart-Grid-Resnet-BBox

A Object Localiztion pipeline based using ResNet as a BBox Regressor(Flipkart GriD 2019-Semifinals)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Flipkart-Grid-Resnet-BBox

Approach Used:

Generally, detection systems classify or localize the object to perform detection. In our system, we have used tried to detect or localize the object without any classification. We have used a regressor similar to ResNet34’s architecture for localization.

Source File:

All of the files are uploaded in a shared folder in google drive with the link as follows: https://drive.google.com/drive/folders/13BOdvE8srZLuOVVE3A8ncWvo91AUAlm6?usp=sharing

Procedure:
1. Install fastai
2. Extract the zip file
3. Move into the extracted folder
4. Using python3 run round3.py with image dataset path as an argument without the last slash (i.e. folder1/folder2, not folder1/folder2/)
5. Round3_final_prediction.csv will be created

About

A Object Localiztion pipeline based using ResNet as a BBox Regressor(Flipkart GriD 2019-Semifinals)


Languages

Language:Jupyter Notebook 100.0%Language:Python 0.0%