WOOD AI is an application made for detecting the defected wood. Wood is a renewable resource for construction of buildings. Construction wood requires strength and due to manual identification of defects in wood, it delays the production time and not feasible for building material reuse.


The main objective is to seek effective methods that can detect different types of wood defects and subsequently classify them. For all industrial machine-based quality control is to reach a minimal sorting error rate with a high speed.

Data Set

Dataset of the wood pieces consist of 4000 images for training of defected and non defected wood.

Data Training Workflow

In the beginning the dataset is cleaned and filtered. CNN is used for training of the defected and non defected wood pieces. The ratio of defected and non defected wood is 80-20%. Trained data is stored in .tfile format for prediction app.

Predicting App Workflow & Deployment

An application with web user interface developed for the wood production companies to predict if their wood is defected or not with automation. The app asks for an image to detect and instantly predict.


Future Applications

Reliable wood type classification, as it provides analytical information. Summarizes worldwide efforts in wood recognition and quality inspection systems. Reduce human efforts as well as make the sorting process much faster.

WOOD AI is a project of IAAC, Institute of Advanced Architecture of Catalonia developed at Master In Advanced Computation For Architecture & Design in 2021/2022 by student: Muhammad Mansoor Awais and Siddhant Chaudhry | Faculty: Oana Taut and Alexksander Mastalski.