Whatsnake

The Problem

Mostly in tropical and subtropical countries, hundreds of thousands of snakes are getting killed by humans daily due to the suspicion that snakes are venomous. Main reasons for such snake kills are inability to identify, differentiate and lack of knowledge on snakes. Most of the time some of the endemic and endangered species of non-venomous snakes are getting killed because these snakes are wrongly identified as venomous snakes.

Our Proposal

We wish to develop a mobile app that detects a particular snake using a camera app and let the user know whether that particular snake is venomous or not (and many other useful info) and how to handle them. With such a solution, helping humans make informed decisions, many reptiles, especially endangered and native snakes in a particular geographic region could be saved. The mobile app will leverage the power of artificial intelligence by building a computer vision tool powered by machine learning. The app will identify the snake, classify whether it is venomous or non-venomous, provide more information on how to carefully handle it, get help from experts and provide relevant first-aid instructions in an unfortunate event of a snake bite.

We Assume that...

We assume that smartphones will be available to a large portion of the population, especially to the rural population where frequent human-snake conflicts are abundant.

We assume that we will be able to gather different images under different lighting conditions, in different angles in order to create a effective machine learning model.

We assume that users will get used to Whatsnake app quickly and properly photograph encountered snakes so that image recognition success rate will be increased.

Constraints to Overcome

Classification of snakes using machine learning algorithm to help users correctly identify snakes whether it's a venomous or non-venomous snake. And help users handle snakes carefully without harming them or getting attacked by them.

Current Work

1. Create a mobile app with necessary features. 2. Initiate data gathering process in order to create an effective machine learning model. 3. Build a image recognition/machine learning model to recognize snakes. 4. Train the ML model with considerable training dataset. 5. Integrate the ML model with the mobile app. 6. Testing and improving the mobile app and ML model.

Current Needs

1. Cloud infrastructure to facilitate application features. 2. Mobile app development resources. 3. Considerable training dataset (photos) of snakes. 4. AI/ML resources.