Disparities in dermatological outcomes may be related to inequities in dermatological education, particularly the lack of darker skin images in educational materials used to train dermatologists and primary care physicians. Thus far, efforts to quantify biases in these materials have been done by hand, which is labor-intensive and impractical for large-scale application. In this work, we aim to develop a standalone online tool that enables a domain-expert to upload a given academic material and obtain distribution of images across different skin tones. This provides first-hand awareness of potential bias existing in these materials and thereby a chance to improve the imbalance. Since all images from a dermatology textbooks or educational materials are not skin related, the first step aims to discard non-skin images. Skin tone estimation needs to also avoid lesions areas of the skin image (e.g., lesion pixels in light skin image might resemble dark skin tone) and unnecessary background and foreground components (e.g., clothes). Results show encouraging performance in detecting skin images and classifying skin tones. The findings confirmed the reports of lack of dark skin images from manual studies of dermatology materials as only an average of 10.15% of dark skin images (compared to light skin images) is found in the four validated textbooks.