System and method for generating virtual pseudo 3D outputs from images

A system for generating a real-time pseudo 3-dimensional (3D) augmented reality image of an object from a 2-dimensional (2D) image of the object, comprising:

• a processor and a user device communicatively connected to the processor, wherein the processor is configured to
• process at least one 2-dimensional (2D) image of an object using a trained machine learning model or a rule-based model to eliminate the background of the object and segment the object from the at least one 2D image, wherein the machine learning model analyses the at least one 2D image on a pixel by pixel basis to segment the object from the at least one 2D image, wherein the pixels that are not associated with the object in the at least one 2D image are determined as background and eliminated by converting into a transparent pixel, wherein the rule-based model processes the at least one 2D image of the object by (i) detecting edges of the object using an edge detection algorithm with in the at least one 2-dimensional (2D) image of the object, and (ii) applying a thresholding algorithm to the edge-detected image to segment the detected edge from the background based on colour and location and (iii) replacing the background of the at least one 2-dimensional (2D) image with a transparent layer, wherein the at least one 2D image is captured using a camera or an image capturing device from at least one angle of view;
• identify at least one key feature and shape of the object in the processed image using a trained machine learning model to align the object image on a stencil/mask used for wrapping the image/texture of the object on a 3D model, wherein the at least one key feature comprises edges, colour, patterns, texture and markers of the object in the at least one 2D image;
• perform bidimensional/UV unwrapping of the 3D model by (i) placing a flat layout of a surface of the object on the UV unwrap of the 3D model, (ii) wrapping the UV on the 3D model, and (iii) generating a pseudo 3D image, wherein the pseudo 3D image is generated for a combination of objects by aligning multiple sets of 2D images together;
• generate a set of consecutive pseudo 3D 360 degree image set of the object based on the generated pseudo 3D image of the object using a frame interpolation techniques that creates new frames between existing frames in a sequence;
• detect movement coordinates of the user device with respect to an image viewer plane using at least one of a change in a camera plane or an inertial measurement unit (IMU) input;
• determine the pseudo 3D 360 degree images to be displayed on the user device by performing real-time interpolation between at least two consecutive pseudo 3D 360 degree images from the pseudo 3D 360 degree image set according to the movement coordinates of the user device, wherein the real-time interpolation provides the pseudo 3D 360 degree image that matches with the 3D image viewed from the at least one angle of view on the user device;
• performing colour correction of the pseudo 3D 360 degree image set to render a realistic visualization of the pseudo 3D 360 object in a current environment by analysing metadata of the at least one 2D image and parameters of the current environment, wherein the metadata is captured during original scanning of the object using the camera or the image capturing device;
• and
• provide subsequent pseudo 3D 360 degree images of the object on the image viewer plane as an augmented reality object in real-time on the user device with respect to the change of the at least one angle of view of the user device from the pseudo 3D 360 image set of the object, wherein the augmented reality object is overlaid on a camera feed or wearable device display of the user device as an interactive object or element.