We are seeking research and industrial partners to jointly develop a FOREST Topic 3 proposal on adaptive visual inspection of wood surfaces across wood species and imaging conditions. Our research concept aims to distinguish natural variations in grain, colour and texture from defects relevant to the intended product use, and to adapt inspection models to new species or production conditions using limited labelled data. We plan to investigate disentangled representation learning and few-shot domain adaptation, with the specific methodology to be developed jointly. The intended outcomes are fewer unnecessary rejects, more reliable defect detection and improved material utilisation in wood processing. We seek partners with expertise in wood science, timber processing, product-specific defect assessment and quality requirements, and companies able to provide representative images, expert annotations and industrial validation settings. Partners with complementary computer vision and domain adaptation expertise are also welcome. Our contribution would focus on AI method development, software implementation and experimental evaluation. The concept is being developed at Maltepe University with PhD researcher Mustafa Yılmaz, who is employed at Gümüşhane University. We are open to joining a compatible consortium or developing a proposal together.

Maltepe University is a university in Istanbul, Türkiye. Our Software Engineering research team works on artificial intelligence, machine learning and software systems. For this concept, we offer expertise in image processing, learning-based visual analysis, software development and experimental evaluation. The research is supervised by Dr. Emre Atlıer Olca and is linked to the doctoral work of Mustafa Yılmaz. We are seeking complementary wood-sector expertise and access to industrial data and validation environments.