Exigent® Target Recognition
Applying Artificial Intelligence to identify targets of interest as early as possible to save lives and protect resources.
Today’s world is consumed with digital images. By leveraging deep learning and long-short-term memory networks to analyze sensor data to automatically assess images, the digital world is able to interact with the physical world using computer vision. Applied to everything from factories to semi-autonomous cars and aircraft, computer vision allows things to run more efficiently and even more safely. Our proprietary AI computer vision technology extends the benefits of AI to safety and security applications, with the ability to autonomously detect targets of interest to protect lives and property.
Exigent® assesses and extracts high-dimensional data from digital images, thereby emulating tasks of the human visual system to transform visual images (the input of the retina) into information and data to aid in making decisions. Once trained on a target of interest, Exigent provides the user the ability to automatically identify targets in the real world.
Computer Vision is about pattern recognition. To train a computer how to understand visual data you feed it images, lots of images – thousands, millions if possible – that have been labeled. Then, you subject those images to algorithms that allow the computer to look for patterns in all the elements that relate to those labels. When it’s finished, the computer will be able to use its experience if fed other unlabeled images to find the ones that are of interest to the user. Once the target of interest is identified, Exigent can send commands to a camera to track the target and continue to provide updates to the user
Exigent® Artificial Intelligence Computer Vision Solution is built around a deep convolutional neural network architecture that exploits spatially-local correlation of features by enforcing a connectivity pattern in the synapses between neurons of adjacent layers. The network is trained using state-of-the-art training algorithms and regularization techniques. The training techniques include Stochastic Gradient Descent, Variable Mini-batch Size, Variable Learning Rate, Dropout and Batch Normalization.
