Intelligent Archiving Of Interior Design Images Using Panorama Picture Sources

Abstract

This Paper Aims To Propose An Approach To Establishing An Intelligent Interior Design Reference Image Database Using 360-Degree Panorama Picture Sources. Building Interior Design Reference Images Are Usually Taken By People In An Ad-Hoc Way And Stored/Used For Various Purposes As Non-Standardized Design Communication Resources. As Several Web/Apps Provide Such Resources, In This Paper We Suggest A Certain Organized Way To Build An Interior Design Image Resource That Can Improve Design Communication Between Architects And Other Stakeholders. Recent Daily-Life Web/Apps Have Collected Architectural Design Images Uploaded By Users, But Stored Images Hardly Used For Professional Design Purposes Due To Issues Of Sufficiency, Quality, Consistency, Copyright, Liability Etc. In This Paper We Focus On The Sufficiency And Quality Issues By Using A Deep Learning Enabled Auto-Classification Of Interior Pictures Extracted From A High-Definition Panorama Picture Source. The Input Image Also Can Be Generated By Photo-Realistic Render Software If It Is A Planned Design. This Paper Describes An Approach To Simplify The Process Of Establishing Such An Archive; 1) Preparing Panorama Image, 2) Auto-Extracting Entire Interior Reference Images, 3) Auto-Classification And/Or Detection Of The Context Represented In Each Picture To Enrich Pictures’ Dataset, And 4) Archiving Image Data With Generated Information To Be Used For Various Design Purposes. This Paper Also Demonstrates An Intelligent Archiving Of Interior Design Reference Images Process Implemented As A Web/App Software Prototype.