An activity classifier system and method that classifies human activities using 2D skeleton data. The system includes a skeleton preprocessor that transforms the 2D skeleton data into transformed skeleton data, the transformed skeleton data comprising scaled, relative joint positions and relative joint velocities. The system also includes a gesture classifier comprising a first recurrent neural network that receives the transformed skeleton data, and is trained to identify the most probable of a plurality of gestures. The system also has an action classifier comprising a second recurrent neural network that receives information from the first recurrent neural networks and is trained to identify the most probable of a plurality of actions.

An accessory assembly for a mobile electronic device includes a platform, having a first side and an opposing second side. The platform includes one or more magnetic elements configured to magnetically attach the platform to a mobile electronic device or case along the first side. The accessory assembly also includes a wallet integrated with or attached to the second side of the platform, where the wallet is configured to hold one or more objects associated with a user of the accessory. The assembly also includes a grip accessory configured to attach to the wallet, where the grip accessory is configured to be handheld in order to support the accessory assembly and mobile electronic device when magnetically attached to the platform.

This disclosure is directed to an activity classifier system, for classifying human activities using 2D skeleton data. The system includes a skeleton preprocessor that transforms the 2D skeleton data into transformed skeleton data, the transformed skeleton data comprising scaled, relative joint positions and relative joint velocities. It also includes a gesture classifier comprising a first recurrent neural network that receives the transformed skeleton data, and is trained to identify the most probable of a plurality of gestures. There is also an action classifier comprising a second recurrent neural network that receives information from the first recurrent neural networks and is trained to identify the most probable of a plurality of actions.

An activity classifier system and method that classifies human activities using 2D skeleton data. The system includes a skeleton preprocessor that transforms the 2D skeleton data into transformed skeleton data, the transformed skeleton data comprising scaled, relative joint positions and relative joint velocities. The system also includes a gesture classifier comprising a first recurrent neural network that receives the transformed skeleton data, and is trained to identify the most probable of a plurality of gestures. The system also has an action classifier comprising a second recurrent neural network that receives information from the first recurrent neural networks and is trained to identify the most probable of a plurality of actions.

In various embodiments, a carry (or carrying) bag is provided that includes an interior, substantially cut-resistant security panel assembly with a matrix of wires secured between or on one or more flexible material layers. Also in various embodiments, the security panel assembly may be positioned intermediate the bag outside wall and a lining of the bag, and in other embodiments, may also take the form of an expansion panel. Second or secondary locking fasteners are also provided to lock first or primary fasteners to or within the carrying bag, to provide security for compartments and pockets. A strap with one or more security cables, and various locking fasteners, may be attached to the carry bag. Methods for forming such security panel assemblies, expansion panels, and carrying straps are also disclosed.

An activity classifier system and method that classifies human activities using 2D skeleton data. The system includes a skeleton preprocessor that transforms the 2D skeleton data into transformed skeleton data, the transformed skeleton data comprising scaled, relative joint positions and relative joint velocities. The system also includes a gesture classifier comprising a first recurrent neural network that receives the transformed skeleton data, and is trained to identify the most probable of a plurality of gestures. The system also has an action classifier comprising a second recurrent neural network that receives information from the first recurrent neural networks and is trained to identify the most probable of a plurality of actions.

A card storage apparatus includes a box body. At least one side of the box body is provided with an accommodating slot, a main control panel, a battery and a sound generator are installed in the accommodating slot. The main control panel is provided with a Bluetooth main control chip and an antenna module, the Bluetooth main control chip is electrically connected to the battery, the sound generator and the antenna module, respectively. The main control panel, the battery and the sound generator are staggered. The card storage apparatus is novel in structure, and functional modules of a Bluetooth tracker are staggered and dispersed, so that a surface space of the card storage apparatus is fully utilized, and the portability thereof is improved.

A wallet is provided with an anti-lost card, so as to facilitate the user to track the wallet by the anti-lost card, the anti-lost card can be detachably or permanently placed in a wallet body of the wallet. A positioning module of the anti-lost card can take advantage of existing network of devices, which work as crowdsourced beacons to ping each other in order to determine the location of the missing wallet, so that compatible electronic devices can use an app to identify the approximate location of the missing wallet.

A cut resistant and highly translucent tote utilizes a multilayer knitted metal fabric having a knitted metal outside layer and a knitted metal inside layer. These two knitted layers may be independent in a portion of the tote, such as in a clear-view portion of the tote, configured to allow clear visibility of items retained in the tote. The independent layers of knitted fabric are highly cut resistant, as they can flex in multiple direction and also move and slide with respect to each other. The tote may have a gusseted base with a base-side sleeve extending at the intersection of the base and sides to allow the tote to be free standing. An opening in the top of the tote may have a closure feature, such as hook-and-loop fastener material.