Image Processing Guidance on HGrid247-3.5
This guide explains how to use HGrid247 to perform image analysis (image processing) on Big Data systems. This process can be run on engines such as:
- Apache Hadoop
- Apache Spark
- Apache Tez
- Apache Flink
In this guide two types of image datasets are used:
- Images of book covers → for image similarity analysis
- Images of faces → for face detection and facial similarity analysis
Dataset Used
1. Images of Book Covers
This dataset contains book cover images .
The purpose of using this dataset is:
- determining image similarity
- finding associations between images
Usage example:
- search for books with similar covers
- book recommendation system
- catalog image deduplication
2. Images of Faces
The second dataset contains images of human faces .
This dataset is used for:
- face detection
- face similarity analysis
- face association
Application examples:
- security system
- face recognition
- person identification in large datasets
Image Processing Module on HGrid247
This guide consists of four main workflows.
The first workflow is used to determine the similarity between images .
Objective:
- compare two images
- calculate the degree of similarity of images
Methods commonly used:
- feature extraction
- vector similarity
The output produced is usually in the form of:
image1 image2 similarity_score
The similarity value is used to determine whether two images are similar or not .
Usage example:
- similar image search
- duplicate image detection
This workflow is used to find relationships between images .
The concept is similar to association rule mining .
Benefit:
- recommendation system
- analysis of the relationship between images
This workflow is a copy of the previous workflow modified for facial data.
Objective:
- comparing two faces
- determine whether two faces come from the same person
Process carried out:
- face detection
- feature extraction
- face similarity calculation
The output produced is usually in the form of:
face1 face2 similarity_score
Used in:
- security system
- face recognition
The final workflow is used to determine:
- interfacial association
- facial similarity in large datasets
