Given that plants and their species are endless in number, creating and maintaining a continuously growing database was a big challenge. To top it, matching and pulling out the right information from a snap was another equally big challenge. Besides ensuring these, we had to ensure that the app is intuitive and engaging enough for the user to go on a learning spree.
We devised a solution based on deep learning because of its ability to perceive information and retain it as knowledge. This empowered our solution to collect and collate information, understand pixel data, tag images and summarize any image of a plant and tell which category it belonged to. Currently, the app has 316,000 species in its database, and can process 71,000 species trained into its algorithm.
We added a host of features to help users ferret information about plants as well as make good use of the information. These include identifying plants by snapshots of its flowers, leaves etc; learning more about type of plants in a specific area, detailed history of when and where an image was taken, browsing to discover exotic plants, typing in details to get the most relevant plant, creating one’s own plant library, and exploring what others are snapping and where they are snapping.
We wanted to make plant search easy and engaging and so paid special attention to the apps layout and operation. Our approach was to give users a logical path to follow and so made the navigation easy and reduced the operations to a couple of swipes. We used leveraged learning to increase the resolution of images, so that there is ample clarity on zooming in. To enhance the overall user experience, we also tested the app thoroughly for loading times and performance speed.
“GetSmartcoders helped us make PlantSnap a comprehensive and easy-to-use app for plant lovers. We thank them for their efforts”