Brick Image Matching App

Brick Image Matching App

Automatically classify and detect the type of brick based on features and similarities.

 

Business Challenge

Automated processing is becoming increasingly popular today. Accuracy, however, is dependent on the experience of the expert. It is challenging to come up with an automated method to analyze diverse brick images for various reasons. Issues should be addressed related to occlusions, lighting, and shadows effects. In addition, it must also be expected that bricks and mortar regions can vary widely in shape, texture, and size, and in some cases, the texture may even vary within the same fragment if it consists of multiple pieces built from different materials over different centuries. Therefore, accurate detection and separation of bricks become necessary.

 

A change by Excelledia

Need to match brickwork whilst on-site or from the office? Our interactive brick image matching app makes it easy to locate the perfect brick. Whether you are looking for bricks that match existing brickwork on a renovation project or bricks that add character to new construction. With this interactive tool – the Brick Image Matching App by Excelledia, architects, specifiers, merchants, housebuilders, developers, and self-builders can select the best brick for their projects. AI algorithms are used to classify images of bricks based on texture size and color and are also used to classify a given input brick image as either a Distressed, Drag faced, Rolled, Smooth or Glazed brick.  Accurate detection and separation of bricks is used in various applications such as stability analysis for civil engineering, brick reconstruction and managing the damage in architectural buildings.

Scope

  • To find similar images to any given brick image:

We developed a similarity model that can find the most similar images to any given input image. The model is created by using the distance metric, calculating the Euclidian pairwise distance between the training images. The model uses these feature vectors and compares them to the input image and sorts for each image a similarity list. The result is visualized for a test brick image that is provided from the folder.

You may choose any brick image from the image folder to be uploaded as the input and when you are ready click upload. As output, the top 3 similar images are displayed along with their distance metric score. You will notice the first image to have a distance metric value to be 0, this is the input image.

  • To classify any uploaded brick image on the basis of its texture:

We developed a classification model that is capable of classifying a given input image as either a Distressed, Drag faced, Rolled, Smooth or Glazed brick. The model is trained/created with the help of training images accumulated from various sources.

You can choose any brick image from the image folder to be uploaded as the input. Once the image is uploaded, the top 3 brick classifications based on texture is displayed along with their percentile score as the output.

 

 

A writer with a flair to convert technical jargons into creative pieces, keeping the reader in mind ensuring to communicate with them through written words and sentences. Always in love with words, after writing, finds peace and needed stress relief from novels and short stories and even a little calming music can do the trick.

Treasa Antony

treasa.antony@excelledia.com

Leave a Reply