Embeddable deep pose estimation

Recent advances in deep learning techniques and improvements in chip manufacturing have enabled the use of convolutional networks in embedded environments. One major challenge for many of these new devices is identifying where in the image an individual may be located and telling what they are doing. Through a combination of techniques an embedable network can be created that can recognize a child in a scene even when mostly obscured.

Data Scientist at knit health working to design and scale deep learning applications in computer vision.