Automatic Detection of Galaxy Type From Datasets of Galaxies Image Based on Image Retrieval Approach

Area of Science:

• Astronomy and Astrophysics
• Computer Science
• Image Processing

Background:

• Galaxy morphology classification is crucial for understanding galaxy evolution.
• Existing methods primarily focus on classification, often neglecting the retrieval of similar galaxy images.
• There is a need for methods that can both classify and find similar galaxies within large astronomical datasets.

Purpose of the Study:

• To develop and evaluate a novel image-retrieval approach for automatic galaxy morphology detection.
• To enhance the capability of identifying galaxy types and retrieving visually similar galaxy images.
• To compare the performance of the proposed method against established optimization algorithms like Particle Swarm Optimization (PSO) and Genetic Algorithm (GA).

Main Methods:

• Feature extraction using shape, color, and texture descriptors from galaxy images.
• Selection of the most relevant features using a binary sine cosine algorithm.
• Computation of image similarity based on extracted and selected features for retrieval.

Main Results:

• The proposed image-retrieval method successfully detects galaxy morphology and identifies similar images.
• Experimental results on the EFIGI catalogue demonstrate superior performance compared to PSO and GA methods.
• The approach effectively handles diverse galaxy types including edge-on spiral, spiral, elliptical, and irregular galaxies.

Conclusions:

• The developed image-retrieval approach offers an effective solution for automatic galaxy morphology analysis.
• This method provides a significant advancement over traditional classification techniques by incorporating similarity retrieval.
• The binary sine cosine algorithm integration proves efficient in feature selection for improved retrieval accuracy.