Assistive Technologies in an abstract term is any technological or scientific discovery or advancement which helps an individual to achieve something more than his/her natural capacity, or in other language helps the person to overcome some barrier like mental, physical, environmental aspects. Specifically, this can mean helping a person to overcome disabilities or improve some ability to improve his/her daily life in context of his day to day operational capabilities. In the recent past there has been a lot of research work for development and advancement of different Assistive Technologies. Hence there has been a surge of need to organize this work in literature manner to help the future researcher and scientists in the research. That’s why there has been increase of parallel work to categorize and organize them. But the majority of the work has been carried out by categorizing them based on the area of the user needs for example, cognitive loss, sensory impairment, visual disabilities etc. This approach which can be termed as “User-need-oriented” categorization has some serious drawbacks. This basically considers each technology or application as a whole which undermines the fact that the same technology can be cross-contextual based on the application. Many a times this type of categorization misses the in-depth aspects of the scientific knowledge used to build the technology and in turn dampens the motivation for research in the assistive technology domain.
To address the critical drawbacks of the most recent trends in literature survey for assistive technology works, the papers takes a drastically new approach. The paper terms it as “task-oriented” approach unlike all the prior work which has been mostly user-need-oriented. It basically divides the final aim of the technologies into different sub problems where each problem are individual tasks. Hence the technology basically is a mixture of different sub technical sub solution which achieves the overall goal. Once the overall goal is divided, each subtask then used to direct to the existing work where that particular task has been used. One major aspects of this type of approach is that this tells that each subtask which solves a particular technical problem can be the component of different user need based goal. For an example, visual recognition technology can be used in assisting visual disable people in indoor environment and can also be used for social robotics in outdoor environments. This type of classification helps to make the work straight forward by highlighting the important issues the technology trying to tackle which can help the future researcher to have a better reading and understanding.Results/Findings
The paper has tried to target two main application contexts one in the context of medical application and the other is the context of social application. To show their approach of task-based classification of assistive technologies, they have tried to classify a set of Computer Vision tasks which already been developed and used to supplement the individual persons capability. For each specific vision task, the paper studies the existing solution and technology which has been used to assist users and also provides future directions in the research both for short and medium terms for the specific task. The different visions techniques they have studied and addressed were visual tracking of objects, human gesture and activity tracking, estimation of the head pose. They also provided some in depth analysis of the different assistive technologies from various perspectives like users, economic, social etc.
The paper gave some pivotal approach for classifying the Assistive Technologies based on tasks rather user needs, which paves a clear path for the future research in this area. The approach is quite novel and unique in terms of the context they have used for the classification. The paper also gives a clear explanation with specifically classification of computer vision assistive technologies based on different technical barriers they try to solve. The organization of paper was quite systematic and easy to read language and structure. Their comprehensive analysis of the approach gives a clear picture to the reader to think beyond the text. Also, the paper gave a concrete future works directions at the end which ends the paper with an optimistic note. One thing which the paper missed is to instead of giving the approach with respect to Computer Vision, they could have given a generic framework which could be extended to multiple technology domains which are the part of Assistive Technologies. Also, the paper misses out to give clear chronological order of past work with proper explanation. Rather they give the past works in a very brief manner which may have undermined the motivation of their work.
Ideas for follow-on work
For future work their work can be made much more generic to establish a framework which can be later extended to be used in any domain of Assistive Technologies. Also, the study results can be improved a lot more in terms of accuracy of data and visualization. One major area of work is to capture the role of Artificial Intelligence in AT especially after the unimaginably fast advancement of Deep Learning technologies. Also, much work is needed in the area of Object detection and tracking, SLAM, noise exclusion etc. how to classify each of them with more and more fine-grained manner to expand the cross-contextual assessment of the Assistive Technologies. Also inspired from this type of classification, there can be new classification frameworks be developed for other technology families like Rescue Robotics, Image Processing, Data Privacy where we can categorize the existing works into different task-based groups. And each task can be part of one specific technology.