This paper presents a computer-aided system that achieves 93.9% accuracy in classifying four common ear conditions (earwax, myringosclerosis, chronic otitis media, normal) using images from a digital otoscope. The system highlights the area of concern on the image and could assist general practitioners in making more accurate ear diagnoses before referral to specialists, especially in regions with limited access to otolaryngologists.

The research received funding from various sources, including CONICYT, Fundación Guillermo Puelma, Fondecyt, Proyecto ICM, which are governmental and non-profit organizations. While the authors declare no competing interests, the funding sources' potential influence on the research direction or interpretation cannot be entirely ruled out. Furthermore, the database creation involved collaboration with the Clinical Hospital from Universidad de Chile where FAC could have conflict of interest with members of the Otolaryngology department. However, there is no evidence or specific details provided to determine the influence of these factors. Thus, further transparency regarding the funding and collaboration would be beneficial in assessing the objectivity of the research.

This study presents a well-designed approach for computer-aided diagnosis of common ear conditions with a high level of validation. The clear methodology, high accuracy and the potential to improve primary care diagnosis warrant a strong rating. However, the limited sample size, potential for bias in data collection, the use of single feature type, lack of external validation and limited comparison with other state-of-the-art methods are weaknesses that require further investigation and prevent it from being a groundbreaking work.