Low-Cost Assistive System for Deaf People Based on Artificial Vision

Hearing-impaired people have trouble communicating with the rest of society because they do not have an effective and inexpensive means of communication. People who learn sign language find it difficult to locate qualified interpreters who can get their message across. This article proposes the development of a software application, using artificial vision and digital image processing to translate the Ecuadorian Sign Language (ESL) into the written language of the Latin alphabet. The objective of this research is to create an economic support system for people with hearing disabilities to improve their autonomy and reduce the cost of medical care. For the creation of the system, digital image processing was performed through four stages: acquisition, preprocessing, the extraction of characteristics, and recognition-translation into the Latin alphabet language. Within the controlled tests conducted in this research, an average processing time of 905.88 ms was obtained, i.e., less than one second for each letter. Regarding the precision in the translation of the system, the efficiency of 97% was obtained.