Abstract

The Ethiopian leather industry predominantly relies on manual inspection by human experts to assess leather quality and identify defects. While this approach has been a longstanding practice, it is inherently limited by human variability, fatigue and subjective judgment, leading to inconsistent grading outcomes and a high incidence of false detections. This study investigates the variability of expert judgements within Ethiopian tanneries and compares the assessments of industry practitioners with academic experts. A longitudinal observational and survey-based methodology was employed, incorporating both qualitative and quantitative analyses. Data were collected from five major tanneries, as well as from the Manufacturing Industry Development Institute (MIDILPR&D) and veterinary professionals. Results revealed significant inconsistencies in defect identification between and within experts, attributable primarily to human cognitive and perceptual limitations. Comparative analysis with international literature indicates similar challenges in other leather-producing countries, reinforcing the need for standardised, automated inspection systems. The findings underscore that automation, complemented by expert training, is essential for improving inspection accuracy, efficiency and fairness in the Ethiopian leather sector.