Watch video presentation: https://www.youtube.com/watch?v=7XPtkbsm-7A
Watch a demo of the video prototype here: https://www.youtube.com/watch?v=i6pqTZRNtjg&feature=youtu.be
One of the most common concerns for child safety is the potential injuries occuring from heavy objects, such as TVs, objects on open shelves, items placed on kitchen countertops, etc. The best way to child proof is to move potentially dangerous items from the reach of children. AI can help identify those potentially dangerous items by detecting their placement in a home and alert the users. Our Deep Learning model serves a great purpose to child proof homes without consulting childproofing experts, particularily, if it is embedded into baby/child monitoring products such as ULO.
Cases
About every 30 minutes, tipped furniture or falling TVs send a child to the emergency room, reports also showed that 349 people were killed between 2000 and 2011 by a falling television, appliance or piece of furniture -- 84 percent of them were kids younger than 9 years old. Falling televisions were more deadly, accounting for about 62 percent of these fatalities.
One prominent hazard from heavy objects is the presence of heavy items, such as electronics, cluttered books, aplliances on top of dressers, large furniture, open shelves etc.
This model can detect this kind of potential hazards and intimate the user.
It can detect the presence of heavy objects such as microwave ovens (not anchored) and other dangerous clutter in kitchen countertops which may be a potentail threat to children.
This picture has open shelf with heavy objects and unmounted TV.
Model
This is a Multi Label Classifier based on Resnet 34 using FastAI.