Abstract: Described herein are systems and methods for training and using a vector embedder to generate vector embeddings for non-natural language data. The method of training the vector embedder includes: receiving the non-natural language data including a plurality of records, each record including a plurality of attributes; grouping the records into a plurality of windows based on a respective subject attribute of the records, each window including a predetermined number of records; sorting the windows in order based on an entropy value of each window; and training a deep learning model to vectorize the non-natural language data by initially training the deep learning model with a set of windows having an entropy value lower than a threshold entropy value to predict one or more attributes of a record in each window based on the other attributes of the records in the window.

Abstract: Methods and devices for performing one step compositing may include receiving at least a first surface from a first image source and a second surface from a second image source. The methods and devices may include generating a composite image with a combined first surface and second surface by contemporaneously applying position corrections to the first surface and the second surface and lens distortion corrections to the first surface and the second surface and blending the first surface and the second surface into respective positions in the composite image. The methods and devices may include transmitting the composite image representing a virtual environment for presentation to a downstream consumer.

Abstract: Methods and devices for performing one step compositing may include receiving at least a first surface from a first image source and a second surface from a second image source. The methods and devices may include generating a composite image with a combined first surface and second surface by contemporaneously applying position corrections to the first surface and the second surface and lens distortion corrections to the first surface and the second surface and blending the first surface and the second surface into respective positions in the composite image. The methods and devices may include transmitting the composite image representing a virtual environment for presentation to a downstream consumer.

Abstract: A multicolored and accented carpet includes a plurality of tufts of color accented fiber including two or more differently colored extruded solution dyed yarns and a base yarn twisted together whereby a plurality of color transitions are provided along the length of the single accented fiber. A cut pile carpet is produced, having the appearance of a space dyed carpet, with the fade and bleach resistance of a solution dyed carpet.

Abstract: The invention provides a drying device for use in drying carpets, rugs after flooding or cleaning by a wet process. In addition, the drying device may be utilised for drying the fabric of dwelling, viz, the floors, walls, ceilings and under floor cavities after flooding. To these ends, the drying device, in one format is made up of opposed sheets (4, 6) of flexible material, e.g. plastic sheeting, co joined along peripheral edge portions (8, 10, 12, 14), one, or both, sheets being provided with a multiplicity of small apertures (24) over the whole, or substantially the whole area thereof The sheets (4, 6) are joined internally by apertured webs (18) extending along the drying device to form channels (20) between the sheets. Provision is made along the peripheral edge portions (8, 10, 12, 14) of the sheets (4, 6) to attach at least one air blower (26) and to attach further drying devices (2a, 2b) using a connector means (30, 38).