VideoKheti extends the reach of agricultural extension beyond traditionally defined boundaries through a mobile interface that puts the low literate farmer in the driving seat.
In July 2014, India had 886 million mobile subscribers (TRAI put the official figure at 933 million users) with the mobile subscription penetration at 70 per cent. Though the rural mobile and internet penetration is still a poor second to urban areas, the pace is picking up. In June 2014, the GSM mobile users grew to over 300 million in rural India. Little wonder that mobile is being viewed as the tool to ensure that the digital India vision percolates down to the grassroots. Much earlier in 2013, Microsoft initiated VideoKheti, a mobile video search system—a research prototype—intended to explore how to provide low literate farmers the professional information that they required. VideoKheti is well attuned to the larger vision of Digital India. It works as a “Youtube for low-literate farmers” by using speech, graphics and touch interaction that helps the low-literate farming communities in rural India find and watch agricultural extension videos in a language or dialect of their choice. Microsoft is currently working to create a more generally-deployed version of VideoKheti.
Wikipedia defines agriculture extension in its simplest form as the application of scientific research and new knowledge to agricultural practices through farmers’ education. In recent years, it has evolved to include communications and learning activities in disciplines as diverse as agricultural marketing, health, etc. Agricultural extension has not only been widely recognised as crucial to the growth of the farming sector, but given the changing nature and the needs of the farmers, there is an added emphasis on developing newer approaches that are region and need specific.
Microsoft Research, in its endeavour to develop systems for helping various underserved communities, has been constantly engaged in finding costeffective technological alternatives. The VideoKheti prototype can make scientific information available to ow-literate farmers in a format that is easy to comprehend. Given the vast geographical spread of rural India and also the fact that farmers reside in remote and inaccessible areas, ICT through mobile technology promises the fastest, farthest and the most accurate reach. VideoKheti can act as the bridge to timely and accurate information for small holding farmers who have little education or means to access other modes of knowledge.
Microsoft Research partnered with Digital Green for the field study of the VideoKheti system to understand its implication at the ground level. Digital Green, an NGO established in 2006, has been engaged in agricultural extension innovation and development of ICT systems in close consultations with the existing people-based extension systems and the farmer community.
The programs are deployed using grassroots social networks for wider outreach. The NGO also uses the local language of the community to facilitate understanding. It has been teaching farmers how to film and then screen their farming best practices for other farming communities. By doing so, they become ‘extensions’ or ‘influencers’ for their peer group of small holding farmers, helping them in the adoption of technologies. In a participatory approach to local content generation, the videos are produced by farmers, are filmed on farmers and are screened for farmers.
The NGO’s reason for adopting visual medium for disseminating knowledge on agricultural practices was guided by the nature of the farming community. Participatory video approach is recognised as an effective communication approach as it allows technological skilling of rural participants and in the case of farmers, it helps them identify problems and develop solutions around them. It also promotes communication between people and groups and is a valuable tool to those working for local empowerment and social change. By sharpening communication skills of rural farmers, it also develops their leadership skills as both are interlinked, according to communication experts. Participatory video is also costeffective compared to other mediums. It can be streamed repeatedly and is a demonstrative medium that can actually show results at close quarters.
In 2012, Digital Green was working with a local manufacturer to develop a mobile device that would function like a camera and a handheld ‘pico’ projector as well. The video model however has been extremely successful. Screened by a local mediator who answers farmers’ queries about the practices that are shown, the videos are commonly stored on an SD card and projected against a wall. In some villages where available,they are also shown on televisions and DVD players. The vision is to replace the pico projectors and TVs with mobiles or tablets by 2017 because they believe that mobiles would expand the reach of the programme by making it more affordable.
VideoKheti seeks to expand the limits of the Digital Green model by addressing two issues: First, farmers require a local mediator to help them view the videos, who then have to explain the demonstrated practices to them. This gives them little opportunity to review the video once the mediator has left the village in case they want to clarify a point or have forgotten a part of the methodology discussed in the video. The second drawback is technical in nature. The existing system makes it an exhaustive exercise for mediators who have to sift through the entire video library every time a video is to be shown. They have to search through the projector interface or manually through the large stack of DVDs, as the system does not offer features to search or browse by crop, season or activity.
VideoKheti tries to address both these issues by providing an easy-touse interface on a single smartphone or tablet. The interface can be used by a mediator as well as by the villagers directly. Designing ICT systems for rural users is pretty challenging. Among other issues, developers must also factor in the low levels of literacy of the target audience, their limited experience in using digital technologies and in a diverse country like India, there is also a multitude of languages and dialects that need to be factored in any mass reach programme. The VideoKheti prototype has been designed to overcome these limitations.
A breakthrough in farmer advisory ICT, VideoKheti can allow users control of content directly, in the language they want to watch the video and also the graphics that they want to view or review. Farmers can find and access agriculture related videos on their personal mobile phones. VideoKheti offers multimodal interaction between the user and the system—the user can use speech input or touch to navigate the system, and the output is a combination of graphics and audio—all at the fingertip of the farmer.
In the current prototype version, a comprehensive collection of 147 videos can be accessed on VideoKheti. Videos have been classified by farming experts according to the features of their content. These features are organized in four levels: crop, crop cycle, type of action, and type of method. A topdown navigation tree based on these four levels has been built and a farmer needs to make a maximum of four choices before reaching a list of videos to play. He chooses a video out of the list. The first level provides the farmer a choice of 22 available crops, shown on two screens that can be navigated using a simple slide gesture. The second level is the choice of the crop cycle (e.g. presowing, harvesting). The third level is the type of action (e.g., field preparation, disease management) and the fourth level is the type of method (e.g., organic, conventional). Some paths lead to the video choice screen in fewer than four levels; for example, the choice of type of method does not make sense for concepts such as irrigation or sowing.
The completely text free interface of VideoKheti is easy to use for semi-literate or even illiterate farmers not used to mobile systems. The choices are organized as a grid where each navigation choice is a grid item. Except for the first navigation level where the various crops are represented by pictures, all graphics are hand-drawn, as recommended in the literature on low-literate users. The interface is based on years of research at Microsoft Research to understand how to build systems that are useful and usable by low literate users.
VideoKheti has an inbuilt system to recognize the dialect of the rural villages. It allows small-vocabulary recognition and is fully automatic and requires a very small amount of training data. The prototype mobile system was tested in 2013 at the field level. Despite the novelty of the system, it was found that users adapted to it fast. Partnering with ground level NGOs like Digital Green, VideoKheti can surely bring a wider network of farmers to learn community-driven farming best practices and reap the benefits on the go. More research to make VideoKheti a true and reliable companion of farmers is on.
