Skip to content

Latest commit

 

History

History
20 lines (16 loc) · 931 Bytes

Readme.md

File metadata and controls

20 lines (16 loc) · 931 Bytes

Image Similarity Detection using Resnet50

Introduction

Given a batch of images, the program tries to find similarity between images using Resnet50 based feature vector extraction.

Usage

python kreas_resnet50.py will compare all the images present in images folder with each other and provide the most similar image for every image.

Pre-Requisites

  • Download Anaconda
  • Make the downloaded shell script executable and install
  • conda -V to check that installation was successfull.
  • conda update conda and conda update anaconda
  • conda update scikit-learn
  • conda install theano
  • conda install -c conda-forge tensorflow
  • pip install keras
  • export MKL_THREADING_LAYER=GNU

Note: More Descriptive Instructions On Installation