Classification of 2D images according to types of 3D arrangement

A computer system comprising:

• a processor coupled to a memory, the memory having stored thereon instructions for performing a computer-implemented process for forming a dataset configured for learning a function, the function being configured to classify 2D images according to predetermined types of 3D arrangement with respect to objects visible in the 2D images, the predetermined types of 3D arrangement including the Manhattan type and/or the Atlanta type, the process being implemented by the processor and causing the processor to be configured to, for each respective type of 3D arrangement:
• obtain information indicative of the respective type of 3D arrangement;
• construct 3D scenes each comprising 3D modeled objects arranged according to the respective type of 3D arrangement,
• generate 2D images each representing a respective 2D perspective of a respective constructed 3D scene where visible 3D modeled objects are among the 3D modeled objects of the respective constructed 3D scene which are arranged according to the respective type of 3D arrangement, and
• add to the dataset training patterns each including a respective generated 2D image and the information indicative of the respective type of 3D arrangement,
• the dataset being applied in machine-learning of a function, the function configured to:
• apply input 2D images, each input 2D image representing:
• a respective 2D perspective of a respective 3D scene where visible 3D modeled objects are arranged according to the Manhattan type, or
• a respective 2D perspective of a respective 3D scene where visible 3D modeled objects are arranged according to the Atlanta type, and
• output, for each input 2D image, information indicative of a respective type of 3D arrangement with respect to objects visible in the input 2D image, the function recognizing the Manhattan type and/or the Atlanta type in the input 2D image.

A computer system comprising:

• a processor coupled to a memory, the memory having stored thereon a dataset formable by implementing a computer-implemented process for forming a dataset configured for learning a function, the function being configured to classify 2D images according to predetermined types of 3D arrangement with respect to objects visible in the 2D images, the predetermined types of 3D arrangement including the Manhattan type and/or the Atlanta type, the process being implemented by the processor and causing the processor to be configured to, for each respective type of 3D arrangement:
• obtain information indicative of the respective type of 3D arrangement;
• construct 3D scenes each comprising 3D modeled objects arranged according to the respective type of 3D arrangement,
• generate 2D images each representing a respective 2D perspective of a respective constructed 3D scene where visible 3D modeled objects are among the 3D modeled objects of the respective constructed 3D scene which are arranged according to the respective type of 3D arrangement, and
• add to the dataset training patterns each including a respective generated 2D image and the information indicative of the respective type of 3D arrangement,
• the dataset being applied in machine-learning of a function, the function configured to:
• apply input 2D images, each input 2D image representing:
• a respective 2D perspective of a respective 3D scene where visible 3D modeled objects are arranged according to the Manhattan type, or
• a respective 2D perspective of a respective 3D scene where visible 3D modeled objects are arranged according to the Atlanta type, and
• output, for each input 2D image, information indicative of a respective type of 3D arrangement with respect to objects visible in the input 2D image, the function recognizing the Manhattan type and/or the Atlanta type in the input 2D image.