A well-developed Computer Vision model requires training based on thousands of previous images that must be correctly annotated. The task of annotating consists of framing and labeling the elements that appear in the images. The Computer Vision model, learning from these images, will be able to detect these elements in other images where elements with similar characteristics appear.
At ASIMOB, today we have Computer Vision models to capture traffic signs, road markings and irregularities in the pavement thanks to these annotations made in a task that, without being very complex, is very arduous.
When we started to train Computer Vision models we received several tips from experts, telling us that we should hire companies located in Asia and Africa, for image annotation. This is how the big tech corporations do it, they told us. The reason for the service relocation suggestion is very simple: it is a cost reason. Annotating is a matter of time, and the cost is quoted in euros per minute of annotating or cents per annotation.
At ASIMOB, that way of doing things did not fit us. Did it make sense to relocate part of our development just because of the cost? Aren’t there people available for that kind of work in our society with high unemployment rates? We wanted to look for local alternatives… and we found them. The solution was given to us by APNABI Autismo Bizkaia. They promoted, through their Occupational Center of the social and labor area, a new way for the inclusion of people with Autism Spectrum Disorder (TEA). 10 people with ASD, who previously worked in manual tasks, considered the proposed tasks interesting and have been carrying out, since the beginning of 2020, the annotation tasks for ASIMOB.
Recently, this collaboration between APNABI and ASIMOB has been awarded the Social Impact Award by the Association of Telecommunications Engineers at TelekoGaua 2022. We have also presented the initiative at the company breakfasts that APNABI holds annually to publicize their projects.
We have been working with APNABI for more than two years and there will be many more, because we are convinced that technology with local social impact can be developed.
Serve this small article to spread it.